The Science of Discworld IV Judgement Da

TEN



* * *



WHERE DID THAT COME FROM?





As the Archchancellor remarks, Discworld runs on narrativium, which tells causality what to do. The same goes for Roundworld, from the outside; that is, as seen by the wizards. But from inside, Roundworld did not have any narrativium until humans evolved and started inventing stories to ‘explain’ all of the puzzling features of the natural world: why it did or did not rain, how a rainbow forms, what causes thunder and lightning, why the Sun rises and sets. We have already seen how these storytelling explanations, often involving heroes, monsters and gods, appeal to a human-centred viewpoint, and how they fail – often dismally – from a universe-centred one.

Many of the greatest questions about causality concern origins. How did plants, animals, the Sun, the Moon, even the world itself, come into existence? We storytelling apes are fascinated by origins. We are not content just to see trees, stones or thunderstorms; we want to know what gave rise to them. We want to see the acorn that makes the oak, to understand the geological story that underlies the stone, and to delineate the electrical genesis of the storm. We want our own special type of narrativium: stories that explain how such things get started, as well as how they work. This wish for simple stories makes us expect simple answers to questions about origins. However, science shows us that our love of stories misleads us. Origins are extremely tricky concepts.

The acorn and the oak have a superficially simple story, which we all understand: plant the acorn, water it, give it light, and it grows into the oak. However, that simple story cloaks a really difficult explanation of an immensely complex development: it is, in fact, much the same account as getting you from an egg. And there’s another complication: not only does the oak come from the acorn: the acorn’s origin is the oak. This is exactly like the chicken and egg cliché. The important question, though, is not ‘which came first?’. That’s a silly question, because they are both part of the repeating system. It’s clear that the chicken is only the egg’s way of making another egg. Before chickens, the same egg lineage used jungle fowl instead to make more eggs; long before that, it used little dinosaurs to make its eggs; and long before that, it used ancient amphibians.

The big problem with ‘turtles all the way down’ as an explanation is not the ludicrous mental image, amusing though that may be. Each turtle is indeed supported by the one beneath. The problem is how and why an entire infinite pile of turtles should exist. What matters in recursive systems is not which part came first, but the origin of the whole system. For eggs and their chickens, that story is mainly an evolutionary one, a sequence of developments that change progressively, so that now we have chickens when previously we had jungle fowl or dinosaurs. In this case, the origins of the system go all the way back to the first eggs, the first multicellular creatures that used embryonic development from eggs as part of their reproductive process. In that same way, the acorn is the modern version of a seed that used to produce early seed plants, and prior to that produced tree-ferns … all the way back to the origins of multicellular plants.

What we mean by ‘immensely complicated development’ also takes a bit of explaining. It’s clear that the acorn doesn’t become the oak tree, any more than the egg that generated you became you. The oak tree is mostly made from carbon dioxide extracted from the air, water from the soil and minerals, including nitrogen, also from the soil. In trees, those ingredients mainly make carbohydrates, cellulose and lignin, along with proteins for the working chemical machinery. The amount of material contributed by the acorn is minuscule. Similarly, almost all of the baby that (in a very restricted sense) became you, was built from a variety of chemicals obtained from your mother through the placenta. The tiny egg contributed very little by way of materials … but an awful lot by way of organisation. The egg functioned to recruit the chemicals that your mother provided, initiating and controlling the succession of stages – blastocyst, embryo, fetus – that led to your birth. Similarly, the acorn is already an embryo, and it has a very complex organisation, beautifully crafted to drive a root down into the soil, to extend leaves up into the air, and to start the business of becoming a tiny oak.

It’s that word ‘becoming’ that we all have trouble with. Jack, on a hospital ethics committee, once had to explain how an embryo → fetus → baby → becomes human. It’s not like switching on a light, he explained; it’s more like painting a picture, or writing a novel. There isn’t one paintbrush-stroke or one word that completes the task; it’s a gradual becoming. ‘That’s fine,’ a lay member of the committee replied, ‘but how far into a pregnancy is it before you have a human being, not just an egg?’ We seem to need to draw lines, even when nature fails to present us with tidily distinguished stages.

So let’s not start with complex development, like acorns and eggs, when thinking about origins. Let’s start with something genuinely simpler: a thunderstorm. Before the storm, there is a time of cool, clear skies, clouds moving with the wind, probably a weather front. What we don’t see, because it’s invisible, is the static electricity building up in the clouds. Clouds are masses of water droplets, billions of tiny spheres of liquid water in a mass of water vapour: a saturated solution of water in air. The droplets and vapour rise to the top of the cloud; then they fall back through the cloud, not quite dropping out as rain, and the cycle repeats. Many do drop out as rain when the storm starts, of course.

Clouds are very active structures, with massive circulations. They look gauzy and simple, but internally they are a mass of water-droplet and ice-particle currents. Each droplet and particle carries a tiny electric charge, and the cloud as a whole also acquires an electric charge, for much the same reason that your nylon underwear acquires an electrical charge opposite to that of your body. So the cloud has the opposite charge to that of the hills it passes over, a clear recipe for trouble. As the charge builds, the electric potential between the cloud and the ground gets bigger. Eventually it becomes big enough for lightning to make its own path between cloud and ground, following a trail of lower-resistance ionised air. Metal spikes sticking up from the ground, or on the top of tall buildings like churches, provide particularly good targets. In the absence of those, a person walking on a hill might be the unlucky Earth-end of a strike.

A thunderstorm seems simpler than an acorn becoming an oak, because it doesn’t need lots of intricate organisation. But even a thunderstorm is not as simple as we tend to imagine: we don’t know how the electric potential builds. There are 16 million thunderstorms each year on Roundworld, but we’re still not really sure how they happen. No wonder we have trouble understanding how an acorn becomes an oak.

As for the origin, the beginnings of a storm, the beginnings of anything … To explain thunderstorms, do we have to explain clouds? The constituents of the atmosphere? Static electricity? The elements of physics and physical chemistry? The origin of anything lies in the interactions of multiple causes. In practice, in order to explain the origin of a storm, or anything else, both the person providing the explanation and the one on the receiving end must have a lot of knowledge in common, covering many different areas. Unfortunately, it may not be present.

You might be an English teacher, an accountant, a housewife, a psychologist, a merchant, a builder, a banker or a student. The chances are that you will not have come across one or more phrases such as ‘saturated solution’ or ‘particle carries a tiny electric charge’. And those phrases are themselves simplifications of concepts with many more associations, and more intellectual depth, than anyone can be expected to generate for themselves.

You might be a biology teacher, a mathematician, or even a science journalist, with a more extensive mental database in such areas. Even so we’d still have difficulty explaining the origin of storms, because we don’t understand it in enough depth. None of us is a meteorologist. And even if we were, we still wouldn’t be able to generate enough depth of understanding for you to be able to say, ‘Ah, yes – I understand that now.’ Jack is an embryologist, and understands eggs and acorns in some depth; he would have the same problem for the same reason, even for those examples. The origin of absolutely anything on Roundworld – of it, off it, all the way up to everything that exists – is a complicated mesh involving enormously many factors that we know very little about.

One way to duck out of this issue is to appeal to divine creation. If you believe in a creator god, you can invoke supernatural intervention to explain the origins of anything, from the universe to thunderstorms. Thor does a great job with his hammer: job done, thunder explained. Or don’t you think so? We don’t find that a very satisfactory explanation, because you then have to explain how the gods came to be, and where their powers came from. Maybe it’s not Thor at all, but Jupiter. Maybe it’s a giant invisible snake thrashing its coils. Maybe it’s an alien spacecraft breaking the sound barrier.

Some quite sophisticated creation stories exist, as mentioned in chapter 4, but none of them are genuine explanations. The same form of words ‘explains’ absolutely anything, and would equally well appear to explain a lot of things that don’t happen at all. If you think the sky is blue because God made it that way, you would be equally happy if it were pink, or yellow with purple stripes, and would offer exactly the same explanation. On the other hand, if you explain the colour of the sky in terms of light being scattered by dust in the upper atmosphere, and discover that the intensity of the scattered light is inversely proportional to the fourth power of the wavelength, then you will understand why short-wavelength blue light will dominate compared to the longer wavelengths of yellow and red. (The fourth power of a small number is very small indeed, and inverse proportionality means that small numbers are more important than big ones – just as one tenth is bigger than one hundredth.) Now you’ve learned something useful and informative, which you can apply to other questions.

However, this type of explanation only goes so far: it doesn’t explain where the dust came from, or more difficult things like why blue light looks blue. If you want complete explanations of anything at all, creation is the way to go. Theology really does have all the answers. Indeed, the myriad religions and creeds on the planet offer a huge choice of answers, any one of which will keep you happy if what you really want is a reason to stop asking why the sky is blue. Attributing it to a deity is just a roundabout way of saying ‘it just is’.

Asimov pointed out that when churches adopted lightning-conductors, they promoted science above theology. Following that way of thinking, we are trying to present scientific – or at least rational – explanations for origins, indeed for many other issues. Ponder Stibbons is the most rational of the wizards, yet even he is fighting an uphill battle. On the whole, though, he’s winning, explaining Roundworld without magic, even though magic – the mechanism behind most Discworld phenomena – is his default viewpoint.

Many, perhaps most, human beings are not rational in their beliefs. Essentially they believe in magic, the supernatural. They are rational in many other respects, but they allow what they want the world to be like to cloud their judgement about what it is like. In the run-up to the American Presidential elections in 2012, several Republican candidates who had previously accepted basic science ended up denying it. A prominent Republican supporter opposed any kind of regulation of the markets on the grounds that this was ‘interfering with God’s plan for the American economy’. More extreme figures on the political right oppose taking steps to mitigate climate change – not because they think it doesn’t exist, but because the quicker we wreck the planet, the sooner Christ’s second coming will happen. Armageddon? Bring it on!

One reason for trying rational approaches first is that most phenomena here on Roundworld have turned out not to be magical. More strongly, many that used to seem magical now make a lot of sense without any appeal to the supernatural: thunder, for instance – though not the American economy, which baffles even economists. So in this book, our explanations of origins will, so far as we can manage, stick to the rational, however complicated it is. But we do wonder whether Christian Scientists, who believe it to be sinful to transplant organs, or even to transfuse blood – because they have been taught that this defies God’s wishes – use lightning-conductors.

Even today, we understand less about thunderstorms than you might imagine.

Two decades ago, astronauts on the space shuttle Atlantis placed a satellite in orbit, the Compton Gamma Ray Observatory. Gamma rays are electromagnetic waves, like light, but of much higher frequency. Since the energy of a photon is proportional to its frequency, that makes them very energetic. CGRO was designed to detect gamma rays from distant neutron stars and remnants of supernovas, and it seemed clear that something was horribly wrong, because the observatory was reporting long bursts of gamma rays, emanating from … the Earth.

This was ridiculous. Gamma rays are produced when electrons and other particles are accelerated in a vacuum. Not in an atmosphere. So something was obviously going wrong with CGRO. Except – it wasn’t. The observatory was functioning perfectly. Somehow, the Earth’s atmosphere was generating gamma rays.

At first, these rays were thought to be generated about 80 kilometres up, well above the clouds. It had just been discovered that strange glowing lights, known as sprites and resembling huge jellyfish, existed at that height. They are thought to be an unexpected effect of lightning in thunderclouds below. At any rate, it seemed clear that sprites must be producing the gamma rays, or at least, associated with them. Theoreticians produced several explanations; the most plausible was that avalanches of electrons produced by lightning were colliding with atoms in the atmosphere, generating both the sprites and the gamma rays. The electrons could move at almost the speed of light and create a chain reaction in which each electron could kick others out of atoms.

From 1996 onward, physicists added bells and whistles to this theory, predicting the energy spectrum of the gamma rays. Data from CGRO fitted these predictions, and confirmed that the rays originated at very high altitudes. It all looked pretty good.

Until 2003.

That year, Joseph Dwyer was in Florida, on the ground, measuring x-rays from lightning, and he observed a huge burst of gamma rays from the storm clouds overhead. The burst had exactly the same energy spectrum as those that were thought to come from much higher. Even so, no one really imagined that the rays that CGRO was detecting came from thunderclouds: they were much too energetic. The energy needed to propel the rays through an atmosphere was too large to be credible.

In 2002 NASA had launched a satellite called RHESSI (Reuven Ramaty High Energy Solar Spectroscopic Imager) to observe gamma rays from the Sun. David Smith hired a student, Liliana Lopez, to look through the data for evidence of gamma rays from the Earth. There was a burst every few days, far more often than CGRO was detecting. This new instrument provided far better information about the energy spectrum, and it showed that these gamma rays had traversed a lot of atmosphere. In fact, they originated at altitudes of about 15-25 kilometres – the tops of typical thunderclouds. As new evidence piled up, it became ever harder to deny that thunderstorms generate gamma rays in huge quantities. Sprites, on the other hand, do not.

How do thunderclouds produce such energetic radiation? The answer is straight out of Star Trek: antimatter. When ordinary matter and antimatter come together, they annihilate each other in a burst of energy – almost total conversion of mass to energy. Antimatter powers Starfleet’s vessels. Its commonest form is the positron, the anti-electron, which is naturally produced by radioactive decay and is routinely used in medical PET scanners (Positron Emission Tomography). However, naturally produced antimatter is rare, and thunderclouds are not renowned for their radioactive atoms. Nevertheless, there is strong evidence that gamma rays from thunderclouds involve positrons.

The idea is this. The electric field inside a cloud is negative at the bottom and positive at the top. This field can sometimes generate runaway electrons with high energies. Being negatively charged, these electrons are repelled by the field at the bottom of the cloud and attracted by that at the top, so they go upwards. They then hit atoms in air molecules and create gamma rays. If such a ray hits another atom, it can produce a positron-electron pair. The electron keeps going upwards, but the positron, having a positive charge, goes downwards, attracted by the field at the bottom of the cloud. On the way down it bumps into an air atom and knocks out new electrons … and the process repeats. Again there is a kind of chain reaction, which spreads sideways, across entire banks of storm clouds.

It’s a bit like a naturally formed laser, in which cascades of photons shuttle to and fro between mirrors, triggering the production of ever more photons as they do so – until they get so energetic that they escape through one of the mirrors. The mirrors are the top and bottom of the cloud, but instead of bouncing photons to and fro, the cloud sends electrons up and positrons down. By 2005 this theory was pretty much firm. The Fermi Gamma-ray space telescope has now detected beams of charged particles, produced by thunderclouds and travelling thousands of miles along the Earth’s magnetic field lines. A substantial proportion is positrons.

This discovery puts thunderstorms in a new light. Not only is Thor’s hammer creating sparks (lightning) and noise (thunder): it is creating antimatter. It’s not the sort of discovery you make by trotting out facile explanations in terms of the supernatural. It depends on repeated scientific questioning of the known ‘facts’.

Even familiar origins lead to new stories as time passes. In its search for rational explanations of origins, science often changes paradigm in response to new evidence or a new idea. The origin of the Earth and the Moon is a good example, with some curious twists. One of them being a short-term failure to change the paradigm in response to new evidence.

In this case, the main problem is too much evidence, rather than too little. We can examine the structure of the Earth, look at the record written in the rocks, and travel to the Moon and bring back specimens. But in some ways this wealth of evidence makes the problem more complicated. What does it all mean? We’re trying to work out what happened, about 4.5 billion years after the event. At that time the universe had already been around for about 9 billion years (according to the Big Bang theory, and even longer according to the main alternatives). In all cosmological theories, the state of the universe gets more complicated as time passes. So by the time our solar system came into being, there was a lot of stuff around.

We have to infer, from what we can observe today, how that stuff aggregated to make the Earth/Moon system. Those observations include data from asteroids, from the Sun and the other planets, and from detailed knowledge of the structure of the Earth and the Moon. (We say ‘the’ Moon, but according to a recent suggestion perhaps there were two or more moons at one stage.) It is clear that there was a time before Earth existed, and then the Earth came into being. The Moon turned up a few hundred million years later. Their origins are intertwined, and we can’t explain one while ignoring the other.

The central problem of the Moon’s origin and Earth’s genesis is that Moon rock is very similar, in subtle chemical detail, to Earth’s mantle. This is the thick layer of rock immediately below the continental and oceanic crust, above the iron core. In particular, the proportions of different isotopes of several elements are the same in rock from either source. This coincidence is too improbable to be compatible with earlier theories of the formation of the Moon, such as the two bodies condensing independently from a primal dust cloud surrounding the Sun, or the Earth’s gravitational field capturing the Moon as it was flying past. George Darwin, one of Charles Darwin’s sons, suggested that the Moon was spun off from a rapidly rotating Earth, but the mechanics – such as energy and angular momentum, a measure of spin – don’t work out correctly. Moreover, the Earth and Moon did not just condense from dust. Astrophysicists and geophysicists now think that the Earth aggregated from many tiny planetesimals, which formed part of a great disc with the burgeoning Sun at its centre. Our telescopes are now good enough to observe several such discs around young suns in neighbouring star systems, and many of these have been found, so that theory seems to hold up.

Between 2000 and the middle of 2012, astrophysicists and geophysicists mostly agreed that the Moon resulted from an enormous collision between an early Earth and an object about the size of Mars. They named it Theia, after the mother of the lunar goddess Selene. This collision vaporised much of the Earth, and nearly all of Theia. Most of the vapour condensed again in lunar orbit, coming together to make the Moon. The rest of it became Earth’s mantle, hence the similarity. The same theory explains the large angular momentum of the Earth/Moon system, a gratifying bonus.

As time passed, problems with the Theia theory began to emerge. It would have produced very high temperatures on Earth’s surface, so pretty much all of the water should have boiled away. This seems incompatible with Earth’s present-day oceans. So extra assumptions were needed to save Theia. Perhaps a few ice asteroids fell on the early Earth and put the water back; perhaps the vaporised water fell back to Earth anyway. However, some very ancient Australian rocks seem to testify to the presence of a lot of water on our world about four billion years ago, which seems to be too soon after the Moon’s origin for such an enormous collision to have occurred.

We described the Theia theory in The Science of Discworld in 1999, but by the second edition in 2002 we were no longer convinced. The biggest problem came from newer computer models of the collision. The first such models, the ones that had established the Theia theory, showed a large chunk of the Earth splashing off; then that chunk split. One part formed the Moon, and the rest fell back to Earth to form the mantle. Theia got mixed up with both of them, but in similar proportions, so anyone could see why both the Moon and the Earth’s mantle had the same composition.

However, the simulations that led to this conclusion took a lot of computing time, and only a few scenarios could be explored. As computers improved, the mathematical models became more sophisticated and their implications could be worked out more quickly and more easily. It turned out that in most of them the bulk of Theia was incorporated into the Moon, while very little went into the mantle.

How can both Moon and mantle be virtually identical, then?

The proposal that was accepted until 2012 was: Theia’s composition must have been very similar to that of the former Earth’s mantle.

This of course is the problem that the whole theory was trying to solve. Why should the compositions be the same? If we can answer that for Theia by declaring ‘they just were’, then why not apply the same reasoning to the Moon? The Theia theory had to assume the same coincidence that it was supposed to explain.

In the second edition of The Science of Discworld, we described this as ‘losing the plot’, an opinion that Ian repeated in Mathematics of Life. This view seems to have been vindicated by the recent (July 2012) discovery of a similar but different scenario by Andreas Reufer and colleagues. This also involves an impactor, but now the body concerned was much larger than Theia (or Mars), and moved much faster. It was a hit-and-run sideswipe rather than a head-on collision. Most of the material that was splashed off came from the Earth, while very little came from the impactor. This new theory agrees with the angular momentum figures, and it predicts that the composition of the Moon and the mantle should be even more similar than had previously been thought. Some supporting evidence for that already exists. A new analysis of Apollo lunar rock samples by Junjun Zhang’s teamfn1 has found that the ratio of isotope titanium-50 (50Ti) to isotope titanium-47 (47Ti) on the Moon is ‘identical to that of the Earth within about four parts per million’.

That’s not the only possible alternative. Matija Ćuk and colleagues have shown that the observed chemistry of Moon rocks could have arisen from a collision if the Earth was spinning much faster at the time – one rotation every few hours. This changes how much rock splashes off and where it comes from. Afterwards, the gravity of the Sun and Moon could have slowed the Earth’s rotation down to its present 24-hour day. Robin Canup has obtained similar results using simulations in which the Earth was spinning only a little faster than it is now, but the impactor was bigger than the Mars-sized body originally suggested.

This is a case where Pan narrans became so committed to an appealing story that it forgot why the story was originally invented. The coincidence that it was supposed to explain faded from view, and a new narrative took over in which the coincidence took back stage. But now the storytelling ape is rethinking the entire story – and this time it is paying proper attention to the plot.

The biggest origin question, philosophically speaking, is that of the universe, which we’ll come to in chapter 18. That aside, the most puzzling origin, a much more personal one, is that of life on Earth.

How did we get here?

Our own inability to create life from scratch, or even to understand how ‘it’ works, makes us imagine that nature had to do something pretty remarkable to produce life. This conviction may be correct, but it could well be misplaced, because a complex world need not be comprehensible in simple terms. Life might be virtually inevitable once the mix of potential ingredients becomes sufficiently rich, without there being some special secret that can be summarised on a postcard. But explaining natural phenomena requires a convincing human-level story. That’s what ‘explain’ means to Pan narrans. However, the stories scientists tell about the origin of life are generally difficult and complex, especially when it comes to filling in details. What happened probably can’t be turned into a story. Even if we could go back and see what happened, what we observed might not make a great deal of sense.

Nevertheless, we can seek stories that provide useful insights.

Most scientific thinking about the origin of life considers two phases: pre-biotic and biotic. Often the problem is simplified further, to inorganic chemistry before life appeared, and organic chemistry afterwards. These are the two great branches of chemistry. The latter concerns itself with the massive and complex molecules that can be formed using lots of atoms of carbon, and the former concerns itself with everything else. And life, as it now exists on Roundworld, makes essential and ubiquitous use of organic chemistry. However, there is no good reason to imagine that the origin of life fits neatly into this convenient but rather arbitrary pair of categories. Organic molecules almost certainly existed before there were organisms to use them. So trying to understand the origin of life as some kind of sudden switch from inorganic chemistry to organic chemistry is a mistake, confusing two different distinctions.

Yes, there was a time before there was life, and a time when life was beginning. But there wasn’t a sudden origin like turning a light on. There was a period, perhaps quite a long one – hundreds of millions of years – of what has been called mesobiosis. This is chemistry, both inorganic and organic, becoming life: the journey, not the starting point or the destination.

A large number of alternative pathways by which life might have originated have been proposed. In the 1980s Jack counted thirty-five plausible theories, and there must be hundreds by now. It is sobering to realise that we may never know which pathway actually happened. Indeed, this is quite likely. The pathway that occurred could well have been one of thousands that we haven’t yet thought of. For some of us, an account which starts in chemistry and finishes in simple biochemistry is sufficient; others will want to see bacterial-grade life produced artificially in the laboratory before being convinced that the sequence of steps can work. Yet others will want to see an artificial elephant, synthesised from chemicals in bottles, and would still insist that someone cheated.

Many of you will be convinced that life is so different from the non-living, even from the freshly dead, that no account of a more or less continuous series of steps can be plausible. In part, this conviction arises from our neurophysiology: we use different areas of our brains for thinking about living or inorganic entities, mice or stones. So it is difficult for us to construct thought-chains leading from stones to mice, or even from school chemistry to ‘germs’. Instead, we come up with concepts like the soul, which makes a clear distinction between our thinking about a living person, and the very different way we think about a dead body.

We’ll summarise some of the plausible accounts of life’s origins, so that you may enjoy the various ideas on offer and the different ways of thinking about the problem that they illustrate. We have written about the origins of life several times in the Science of Discworld series, so we will try to make this account a little different. The virus story at the end of this section, for instance, is quite new. It was sitting quietly behind the scenes around 2000, but in 2009 a review paper by Harald Brüssow opened it up for discussion. To put it in context, we need to look at some of the earlier proposals.

The most important early experiment was that of Stanley Miller, working in Harold Urey’s laboratory in the 1950s. He imitated the effects of lightning on a reasonable approximation to Earth’s early atmosphere: ammonia, carbon dioxide, methane and water vapour. Initially, he got several noxious gases, like cyanide and formaldehyde, both notable poisons; this encouraged him, because ‘poison’ is not an inherent property – it describes an effect on living organisms. Most gases don’t get involved with life at all. Further runs of the experiment produced amino acids, some of the most important chemicals for life, because they aggregate into proteins. He came up with a variety of other small organic molecules as well.

Understanding how these molecules came into existence would be very complicated, but the experiment shows that nature can achieve the result without making any special effort. There is no reason to suppose that anything beyond standard chemistry, obeying the usual physical and chemical rules, is involved in Miller’s experiments. We can tell plausible chemical tales about reasonable ways for the atoms and molecules to combine and change. It happens all the time; this is why the subject ‘chemistry’ exists. Reasonably detailed models would capture the main steps – but the reality is almost certainly more complicated than those models. This is an important principle: what seems complicated to us may be easy for nature.

Workers repeating this experiment with different reasonable atmospheres have obtained many other organic compounds, such as sugars, and even the bases that link up to make DNA and RNA, key molecules of terrestrial life. We’ve already mentioned DNA and its double helix, and in any case it’s very well known nowadays. RNA, which stands for ‘ribonucleic acid’, is less well known: it is like DNA, but simpler. With a few exceptions, RNA forms a single strand instead of two intertwined ones. Specific forms of RNA play vital roles in the development of an organism.

These two molecules could easily have been present in the early seas on our planet; indeed, they were probably unavoidable. In addition, we now know that meteorites contain many of these simple organic compounds; indeed, they can form in empty space. So that’s another sensible source of organic chemicals. In short, small organic molecules were around, in quantity, for reasons that have nothing to do with living organisms.

This simple chemistry, though a promising start, isn’t enough. The key molecules in organisms are far more complicated, involving vastly more atoms arranged in fairly specific ways. Graham Cairns-Smith suggested that clay molecules would be ideal catalysts for turning simple organic compounds into polymers of the kinds found in living things: linking amino acids into peptides and proteins, and possibly linking bases with phosphorus and sugars to form short strings of nucleic acids including RNA and DNA. Again, nothing beyond standard chemistry is required to achieve this, and the processes do not involve living creatures. So it would be surprising if there were not many polymers in the early seas. Getting complex molecules is not a problem. We may have trouble coping with their complexity, but nature just follows the rules; from this, a sort of complexity unavoidably follows.

However, polymers aren’t alive. They don’t reproduce, or even replicate, except in very special situations. (Replication is the making of exact copies; reproduction is the making of inexact copies which nevertheless can themselves reproduce, which is more flexible, but even harder to understand.) Replication or reproduction seem to require not just complexity, but organised complexity, and it’s difficult to see where the organisation can come from. However, some of these special situations can occur entirely naturally with certain clays, which themselves exhibit replication. In a watery environment, little slabs of clay can make stacks of almost identical copies, without any help.

Since the late 1990s many things have changed. At that time, in The Science of Discworld, we paid particular attention to the ideas of Gunther Wächtershäuser. His proposal differed from the by then conventional Miller-Urey primeval soup, which spontaneously produced replicating nucleic acids, the first heredity. Instead, Wächtershäuser proposed that the first thing to happen was metabolism: working biochemistry. He suggested that this might have occurred where there was plenty of sulphur, iron oxide and iron sulphide, plus a suitable source of heat to drive the chemistry. One possible location that possesses these ingredients is an undersea hydrothermal vent, known as a black smoker, where molten rock from the mantle makes its way to the surface through cracks where the ocean floor is spreading. Less dramatically, underwater volcanic vents do the same. Using this kind of iron-oxygen-sulphur chemistry, Wächtershäuser came up with a set of chemical reactions that closely mimicked the Krebs cycle, a central biochemical system in nearly all living things.

In laboratory experiments, his scenario performed reasonably well, though not perfectly. So the theory of the origin of life turned into a kind of primeval pizza, with molecules dotted around on a surface, rather than a primeval soup sloshing around in pools or the open sea. In 1999 we liked this idea because it was different from heredity-first systems: we couldn’t see why they would necessarily replicate – what was in it for them. Moreover, Wächtershäuser was a lawyer as well as a biochemist, and it’s unusual to get good original scientific ideas from a lawyer.

However, since then a different idea, the RNA world, has really taken off. RNA and DNA are both nucleic acids, so named because they are found in the nuclei of cells. There are many other kinds of nucleic acid; some are much simpler than DNA and RNA, and some are much more complicated. Both are long chains formed by joining together four smaller molecular units, called nucleotides. Nucleotides are combinations of bases, which in turn are specific molecules that look like complicated amino acids, linked together by sugars and phosphate. Does that help? We thought not. You can look up all the details in many sources, but for present purposes we just need convenient words to keep straight which bits we’re talking about.

The great trick that nucleic acids exploit is their ability to form double-chains, each half encoding the same ‘information’ in related ways. The DNA code letters, the four bases, come in two associated pairs, and the sequence of bases on one chain consists of the partners of the bases on the other chain. This makes possible the key feature of these pairings: each chain determines what happens in the other chain. If they split apart, and each chain acquires a new partner, by sticking on the complementary bases … lo and behold: originally we had one double-chain, and now we have two of them, each identical to the first. The molecule not only can replicate: it does, given enough unattached bases to play with. It would be hard to stop it.

RNA has other tricks. It can function as an enzyme, a biological catalyst; it can even be the catalyst for its own replication. (A catalyst is a molecule that promotes a chemical reaction without being used up: it gets involved, helps things along and then ducks back out.) And it can also catalyse other chemical reactions that are useful to life. It’s a universal fix-it molecule for living organisms. If it were possible to explain how RNA could appear in the absence of life, it would constitute a wonderfully useful step from non-living chemistry towards a primitive kind of life form. Unfortunately, it turned out to be very difficult to see how RNA could turn up in the primeval soup without any assistance. For many years, the RNA world theory was missing one of its most vital features.

This is no longer an obstacle. In recent years, many different solutions to this problem have been found, including several that work experimentally as well as theoretically. The chains of bases involved were initially fairly short – a chain of six is easy, but now there can be fifty or more bases in a chain. This is getting close to the length found in real biological enzymes, which usually have 100-250 bases. So there is real hope that long RNA chains must have been present in that ancient soup. More plausibly still, fatty membranes, which closely resemble cell membranes, have been synthesised in circumstances very similar to those that are thought to have existed on the primitive Earth, and RNA gets linked to these in useful ways. Moreover, it has recently been suggested that the RNA chains could repeatedly be broken apart – unzipped – by high temperatures in black smokers and reassembled at lower temperature in cycling convection currents. This is a lovely idea, exactly like the way DNA is multiplied in systems that analyse its sequence using the polymerase chain reaction, where alternating high and low temperatures break the chains apart and then permit them to build complementary chains, repeatedly doubling the number of copies. RNA could be replicated by this natural physico-chemical process.

For these reasons and many others, the RNA world has now become a respectable image for the earliest stages of life on Earth. It may not be what actually happened, but it provides a plausible scenario. And even if life did not arise in that way, the RNA world shows that there is no compulsion to invoke the supernatural. In the primitive seas, probably around smokers but perhaps on tidal beaches where pools were concentrated – and irradiated – in sunlight, and diluted when the tide came in, or under the influence of volcanoes or earthquakes, RNA strings were growing and being copied.

The copying process wasn’t always perfectly accurate, but that was a positive advantage, because it led, without any special interference, to diversity. If random variation of this kind could be coupled to some kind of selection, favouring sequences with specific features, then RNA could – had to – evolve. And selection wasn’t an issue; the big problem would have been to prevent it. As special sequences with particular properties appeared, competition between them for free nucleotides, and probably for interactions with particular fatty membranes, eliminated some sequences and encouraged others to proliferate. This led to longer chains with even more special properties.

Natural selection had begun … and the system was becoming alive.

In this view, not only does evolution by natural selection explain how life diversified: it is part and parcel of what brought it into being in the first place. Copying errors, if they occur, though not too often, can be creative, in the context of a sufficiently rich system.

The RNA world is not the only game in town. The latest proposals for the origin of life hinge upon viruses. Viruses are long DNA or RNA chains, usually surrounded by a protein coat that contrives to inject them into other organisms, especially bacteria and animal and plant cells. Most viruses rely on the DNA/RNA copying machinery of the organism they infect to replicate them. Then the new copies are usually sprayed out into the environment when that cell, or the organism, dies.

Since the work of Carl Woese in 1977, taxonomists – scientists who classify life into its innumerable related forms – have recognised three basic kinds of life form, the largest and earliest branches of the tree of life. These ‘domains’ comprise bacteria, archaea and eukaryotes. Creatures in the first two domains are superficially similar, being micro-organisms, but each domain had a very distinctive evolutionary history. Archaea may trace back to the earliest organisms of them all; many live in strange and unusual environments: very hot, very cold, lots of salt. Bacteria, you know about. Both types of organism are prokaryotes, meaning that their cells do not clump their genetic material together inside a nucleus, but string it from the cell wall, or let it float around as closed loops called plasmids.

The third domain, eukaryotes, is characterised by having cells with nuclei. It includes all complex ‘multicellular’ creatures, from insects and worms up to elephants and whales. And humans, of course. It also includes many single-celled organisms. RNA sequences imply that the first big split in the tree of life occurred when bacteria branched away from ancestral archaea. Then that branch split into archaea and eukaryotes. So we are more closely related to archaea than we are to bacteria.

Viruses are not part of that scheme, and it is controversial whether they are a form of life because most of them can’t reproduce unaided. It used to be thought that viruses had two different origins. Some were wild genes that escaped from their genomes and made a living by parasitising other creatures and hijacking their gene-copying equipment. The others were desperately reduced bacteria or archaea; in fact, they were reduced so far in their pursuit of a parasitic lifestyle that all they had left were their genes. Occasionally it was supposed by lay people, physicists or maverick biologists – who should have known better – that, being so simple, they were relics of ancient life. This incorrect line of thinking stems from the same, mistaken, principle as the one that considers Amoeba to be ancestral because it looks simple. Actually, there are many kinds of amoeba, and some have 240 chromosomes, bodies in the cell that contain the genes. We have a mere 46 chromosomes. So in that sense an amoeba is more complex than we are. Why so many? Because it takes a lot of organisation, in a very small space, to get all of an amoeba’s functions to work.

Brüssow, in 2009, wrote a review called ‘The not so universal tree of life or the place of viruses in the living world’.fn2 In it he pointed out that Darwin’s tree of life, a beautiful idea that derives from a sketch in The Origin of Species and has become iconic, gets very scrambled around its roots because of a process called horizontal gene transfer. Bacteria, archaea and viruses swap genes with gay abandon, and they can also insert them into the genomes of higher animals, or cut them out. So a gene in one type of bacterium might have come from another type of bacterium altogether, or from an archaean, or even from an animal or a plant.

The major agents of this swapping are viruses, and there are lots of them on the planet, probably ten times as many virus particles as all other forms of life added together. Now, it might appear that with all this swapping, it would be virtually impossible to work out lineages, the heredity of individual bacteria. Even more so, it would seem impossible to work out the lineages of the viruses doing the swapping. Surprisingly, this is not the case; well, not altogether. There are clues in the order in which specific genes appear in many viruses, and there are useful clues as to which organisms the viruses parasitise. Some parasitise both bacteria and archaeans, suggesting strongly that they have been doing so since before these groups diverged. Moreover, these are viruses with RNA genetics. So Brüssow proposes quite convincingly that these particular viruses may be relics of a former RNA world. Going further: infection by ancient DNA viruses could have imported DNA into the heredity of all of the familiar creatures whose genomes we now make such a fuss about. So some of the mavericks, and physicists, might have been right all along, even if it was for the wrong reasons.

If that is the case, we need to look with new eyes at all the ways in which RNA is involved in modern life forms. According to the standard story, which has not changed for some time, RNA serves as a humble messenger that carries the all-important DNA sequence to the ribosomes, huge molecular structures in which proteins are assembled. There are also small RNAs that transfer amino acids to the ribosomes for protein assembly. Ribosomes in turn are made of several kinds of RNA, and several people have suggested that they are the central mechanism of protein construction in cells, their most important function.

This story may soon have to change.

There has been a revolution in nucleic acid biology over the last ten years, and almost all of it has been about RNA. Messenger RNA and transfer RNA are merely the most prosaic jobs that RNA performs in cells. But RNA does many more important – perhaps having said ‘prosaic’ we should now say ‘poetic’ – jobs too. When DNA was considered the most important molecule in the cell, and protein construction the most important function (many textbooks still think so), strings of DNA that specified proteins by transcribing messenger RNA were called genes. The strings of DNA upstream and downstream, which did not specify proteins, were mostly thought to be ‘junk DNA’ of no importance to the organism. Junk DNA was just there as an accidental by-product of past history, but because it didn’t cost much to replicate it, there was no evolutionary reason to eliminate it.

Indeed, there are plenty of remnants of old genes, and quite a lot of sequences from ancient viral attacks, which really might be junk. However, it turns out that although it doesn’t specify proteins, nearly all DNA between genes does transcribe RNA molecules. These RNA molecules form the main control system of cells: they control which genes are activated and when, and how long different messenger RNAs last before destruction. In bacteria they also control genes, but a subset of them protects the bacterial cell against attack by viruses. This is a simple kind of immune system. So DNA may be the soloist, but RNA is the orchestra.

With that established, we can return to ribosomes, the molecular factories that assemble proteins. They are tiny particles, mostly of RNA. In bacteria, archaea, animals, plants and fungi, every cell has its own complement of ribosomes; moreover, much the same RNA occurs, though with different proteins, right across the span of life.

Marcello Barbieri is a leading exponent of biosemiotics, a relatively new science concerned with the codes of life. You have probably heard of the genetic code, the way in which triplets of DNA nucleotides are turned into different amino acids in proteins – by the ribosome. Barbieri has pointed out that there are hundreds of other codes; they range from insulin anchoring itself to receptors on the cell surface and causing different effects in the cell, to a smell (technically a pheromone) in male mouse urine that affects the oestrous cycle of female mice. All such effects are the results of translating one chemical language – hormones, pheromones – into a different language – physiological effects. So the genetic code is not alone: there are codes everywhere in biology. From this viewpoint, the crucial element in protein formation is not the DNA that prescribes it, or the messenger RNA that transmits the prescription: it’s the ribosome. Which, to complete the analogy, is the pharmacist that makes up the prescription.

It also seems clear that this very ancient piece of machinery, so central to all living function, pre-dates the bacteria/archaea split, so it probably derives fairly directly from the RNA world. Something back then formed a relationship, a translation, probably from nucleic acid to protein. The ancestor of today’s ribosomes, probably not very different from today’s range of RNA structures, did the trick. So at the beginning of life, we find the translation of one kind of chemistry into another, by a structure that has come down to us almost unchanged.

Before the ribosome, there was just chemistry. Complicated chemistry, to be sure, but complication alone isn’t quite the point. What matters is complexity, which in this context means ‘organised complication’. Every cook knows that heating sugar with fats, two fairly simple chemical substances, produces caramel. Caramel is enormously complicated on a chemical level. It includes innumerable different molecules, each of which has thousands of atoms. The molecular structure of caramel is far more complicated than most of the molecules you’re using to read this page. But caramel doesn’t do much, aside from tasting good, so mere complication isn’t enough to make interesting new things happen. Similarly, mixing dilute solutions of amino acids, sugars, bases and so on with particular clays generates long, very complicated, polymers. But, like caramel, they’re not very interesting. However, as soon as transactions between those molecules came about, via the earliest ribosomes, complexity took over from complication.

Here, ‘complexity’ refers to organised complication. In a complicated system, such as a car, the individual bits – brakes, steering wheel, engine – behave in much the same way outside the system as they do when they’re part of it. Mostly, they just sit there unless they’re pushed, or pulled, or operated, by something else. But you, a fly, or an amoeba, are different. Their components behave differently when they are part of the system compared to what they do on their own. The parts interact more closely, changing their nature in the context of the system.

A bridge linking an island to the mainland is a complex system in this sense. In order to do its job, it doesn’t much matter what the bridge is made of: it could be rope, steel or concrete. It could even be made of nothing (or air) if it’s a tunnel. The important property is not what it’s made from, but that it links the two ends effectively. That linkage is an emergent property of the bridge. That is, it’s not inherent in any of the materials used. It arises because of their relationships to each other and to the local geography. Moreover, once the bridge is in place, the local geographical function is changed. The river that the bridge spans is no longer an obstacle to vehicles, even though they can’t float or travel underwater. Crucially, you won’t understand how that change occurs by studying the materials that made the bridge.

When the two ends of the bridge are linked, and only then, the local geography changes dramatically. So the real origin of a bridge occurs when the ends are linked. For some purposes, this is when the first rope crosses the divide; for other purposes it’s when the first car makes the crossing; for yet other purposes it’s when the Customs Office is set up.

Similarly, a ribosome in a cell is very different from an isolated one. It has a specific but complex job to do, reading messenger RNA and constructing proteins according to the genetic code. We wonder whether the chemical transactions made possible by early ribosomes in effect constructed bridges between several different kinds of chemistry, providing energy and materials for the ribosome to replicate itself. It’s mostly RNA, after all.

Indeed, if we had to point to a single innovation that marked biotic from pre-biotic, it would be the ribosome, the translator supreme. Barbieri thinks the ribosome is central to life, and so do we. DNA is simply the rather prosaic, boring text. The ribosome is the orator; the other RNAs are the poetry. Once the ribosome emerged, the future became a living future, and in many ways this step marks the true origin of life.

Most origins also involve more subtle forms of emergence: the beginning of a storm, the acorn’s origin as a bud on the oak, the origin of the Earth. Each of these origins is a quantitative-to-qualitative transformation, an emergent event that localises a real beginning. The first stroke of lightning, the first pair of leaves, the generation of heat that melts the core inside the Earth’s mantle: these are emergent events that can label beginnings of new structures. The ‘becoming’ has divided into two issues, before and after the emergence.

If a phenomenon is emergent, it transcends all that has gone before. It does something that its bits and pieces could not have done on their own, or partially assembled, or assembled with some extra scaffolding that gets in the way. This transition is often the best stab we can make at assigning an origin. An emergent phenomenon does not originate in the bits and pieces that led to it: it originates when it emerges.

The emergence of the first lightning strike marks the beginning of the storm. The cell divisions that mark the acorn’s difference from the other buds around it are the emergent oak. The cell divisions and relationships that promoted the egg that later became you orchestrated the emergent event that began you. The universe is complicated because emergent events – quantitative differences becoming qualitative differences – have occurred so many times. Bridges like ribosomes have been built, and the Moon now circles the Earth.

These links have joined separate events into a web of causality that is the most notable property of the world around us. A story, however, is not a web. It has a linear structure, because both speech and writing proceed one word at a time. Even hypertext, used on the internet, is determined by a linear programme written in hypertext mark-up language (html). And that is why storytelling – human narrativium – finds origins to be so difficult and puzzling, and sometimes looks for simplicities where none exist.

fn1 Junjun Zhang, Nicolas Dauphas, Andrew M. Davis, Inigo Leya and Alexei Fedkin, The proto-Earth as a significant source of lunar material, Nature Geoscience 5 (2012) 251-255.

fn2 Harald Brüssow, The not so universal tree of life or the place of viruses in the living world, Philosophical Transactions of the Royal Society of London B364 (2009) 2263-2274.