The way houses are built is called ‘top down’. In this sense of ‘top’, the architect’s plan is at the top. The architect draws a set of detailed plans: a plan with precise dimensions for each room, detailed instructions as to what each wall is made of, how it is to be finished, where the water pipes and electric cables are to run, exactly where each door and window is to be, the precise location of every chimney and fireplace and supporting lintel. These plans are passed down to bricklayers and carpenters and plumbers, who take them and follow them meticulously. That’s top-down building, with the architect – or rather, the architect’s plans – directing the whole procedure from the top. That’s ‘blueprint building’.
Bottom-up building is very different. The best example I know is a termite mound. Look at this page and be amazed. Daniel Dennett made a fascinating comparison to illustrate the distinction between bottom-up and top-down design – and the potential similarity, and complexity, of the results. On the right of this pair of illustrations is La Sagrada Família, a beautiful church in Barcelona. On the left is a termite mound, photographed by Fiona Stewart in Iron Range National Park in Australia. It’s a mud nest built by a colony of termites. Actually, most of the nest is underground. The ‘church’ on the surface is an elaborate set of chimneys whose purpose is ventilation and air conditioning of the underground nest.
The resemblance is almost spooky. But the Barcelona church was designed, down to the last detail, using blueprints. Designed by the famous Catalan architect Antoni Gaudí (1852–1926). Nobody and nothing, not even DNA, designed the termite mound. Individual termite workers built it by following simple rules. No termite has the foggiest idea of what a termite mound should look like. None of them has anything like a picture or plan of a mud church in its brain or in its DNA. There never was a picture, or a blueprint, or a design for a termite mound, anywhere. Each individual termite just follows a set of simple rules, on its own, with no idea of what the other termites are doing, and no idea of what the finished building will be like.
I don’t know exactly what those rules are, but this is the kind of thing I mean by a simple rule: ‘If you come across a pointy cone of mud, stick another dollop of mud on it.’ Social insects make use of chemicals – coded smells called pheromones – as an important system of communication. So the rules followed by individual worker termites when building a tower might depend on whether a particular piece of the edifice smells like ‘this pheromone’ or like ‘that pheromone’. When ‘design’ emerges from the obeying of simple rules, where there is no overall plan in existence, anywhere, it is called ‘bottom-up’, as opposed to ‘top-down’, design.
This page shows another beautiful example of bottom-up ‘design’, starlings flocking in vast numbers in winter. In this case what’s being ‘designed’ is behaviour, a sort of aerial ballet rather than a building. So, instead of saying ‘There’s no architect’ I’m going to say ‘There’s no choreographer’. Nobody knows quite why they do it, but as evening approaches the birds congregate in huge flocks which can contain thousands of individuals. They fly together, fast and with such precise coordination that they don’t collide, wheeling and turning together as though following direction from a master bird. A flock of starlings moves like a single animal. The ‘animal’ even has a distinct and definite edge. You should really look at some of the breathtaking movies of this wonder of the world. Search YouTube for ‘Starling winter flocks’.
While you watch these flocks wheeling and soaring and diving, as though this huge conglomeration of birds were one giant animal, you can’t help feeling there must be a master flight-coordinator, perhaps a single boss bird communicating with the others by telepathy: ‘Turn left now, wheel up and around, now swing to the right…’ It looks totally top-down. But it isn’t. There’s no director, no conductor, no architect, no boss bird. In a way now coming to be understood, all the individual birds, each one following bottom-up rules, together produce an effect which looks top-down. It’s like the termites again, but on a faster time-scale. And what they produce is not a mud church but a superb aerial ballet – with no choreographer.
The power of this bottom-up non-choreography was beautifully demonstrated by a clever computer programmer called Craig Reynolds. He wrote a program called Boids to simulate flocking birds. You might think that Reynolds programmed the whole movement pattern of the entire flock. He didn’t: that would be top-down programming. Instead, his bottom-up program worked like this. He put a lot of effort into programming just one bird, with rules such as: ‘Keep an eye on your neighbouring birds. If a neighbour does so and so, you must do such and such.’ Having perfected the rules for his one bird, he then ‘cloned’ it: made dozens of copies of the one bird and ‘released’ them all into the computer. Then he watched how the whole flock behaved. The boids flocked very much like real birds. This page shows a yet more beautiful simulation, building on Reynolds’s one, programmed by Jill Fantauzza for the San Francisco Exploratorium.
The important point is that Reynolds didn’t program at the flock level. He programmed at the level of the individual bird. Flock behaviour emerged as a consequence. Such ‘bottom-up’ programming is also how embryology works, with individual cells in an embryo playing the role of individual birds in a flock. Embryological development involves a lot of movement of cells, with membranes and sheets of tissue folding and caving in dynamically. So, as with the flying starlings, we are talking about ‘no choreographer’ as well as ‘no architect’.
Embryologists work on how DNA builds a baby. Quite a lot is now known, but I’m not going to discuss it in detail. It would take a whole book, and it’s not what this book is about. For our purposes, we just need to understand that embryonic development, the process by which bodies are built, is a bottom-up process. Like the way termite mounds are built, or flocks of starlings are coordinated. There is no blueprint. Instead, every cell in the developing embryo follows its own little local rules, like individual termites building a mud cathedral or individual starlings in a wheeling flock.
I’ll go just a little bit further, into very early embryo life, to show how these bottom-up rules work. The fertilized egg, as you know, is a single cell. A big one. It splits into two. Then each of the two splits, to make four. Then those four split to make eight, and so on. After each split, the total size remains the same as the original fertilized egg. The same material is divided among two, four, eight, sixteen cells and so on, forming a solid ball. By the time the number of cells has reached a hundred or so, they have formed themselves (following local bottom-up rules) into a hollow ball, called the blastula. Once again, the size of the blastula is about the same as that of the original fertilized egg cell, and the cells themselves are now very small. The outside of the ball is a wall of cells.
The number of cells goes on increasing, as the cells split again and again. But the ball doesn’t become bigger. Instead, again by each cell following local rules, part of the wall becomes dented in towards the middle of the ball. Eventually, the denting has gone so far that the ball is lined by two layers of cells instead of only one. The double-walled ball is called the gastrula, and the process of making it is called gastrulation.
Admittedly a gastrula is not very complicated, and it doesn’t look at all like a baby. But I think you can see how bottom-up rules followed by each cell, working on its own, could form the gastrula – by expanding the wall of the blastula and causing it to dent in to make the double-walled gastrula. And it’s bottom-up rules like this that continue, working locally all over the embryo, to change the shape so that it steadily becomes more like a baby.
After gastrulation, another somewhat similar ‘denting’ process occurs. In this one, called ‘neurulation’, the denting ends up by pinching off a hollow tube, which is destined eventually to turn into the main nerve cord (the one that in each of us runs all the way down the back inside the spine). Again, the denting in neurulation works by individual cells following bottom-up local rule
s. The picture here shows how the nerve tube is made, first by ‘denting’ and then by a ‘pinching off’ of the dented part. The details are different from gastrulation. But the same principle of bottom-up local rules is at work.
You remember how Craig Reynolds wrote a computer simulation of a flock of birds – ‘Boids’ – by programming the behaviour of just one ‘boid’. He then made lots of copies of his one ‘boid’ and watched how they behaved together. They formed a flying, wheeling flock, just like real birds. Reynolds never programmed flock behaviour. Flock behaviour emerged, bottom-up, as a consequence of individual boids following local rules. Well, a mathematical biologist called George Oster did the same kind of thing, but with cells in an embryo instead of boids. He wrote a computer program to simulate the behaviour of a single cell. To do this he used lots of details that biologists already knew about single cells. Really quite complicated details, because cells are complicated things. But the important point is this. As with the boids, Oster didn’t program an embryo. Just a single cell. Including the tendency to divide, which is one of the important things cells do. But cells do other things too, and Oster programmed them into his single cell, as well. He then let it divide on the computer screen, to see what happened.
As the cell divided, each copy inherited the same properties and the same behaviour as the original cell. So it was like Craig Reynolds cloning up lots of copies of his single boid, to see how they would behave in a flock. And, just as Reynolds’s boids flocked like starlings, Oster’s cells…well, just look at the diagram on this page to see what they did. And compare it with the picture of real neurulation, above. Of course, the two are not exactly the same. Nor were Reynolds’s flocking boids exactly the same as real flocking starlings. In both cases, all I’m trying to do is show you the power of bottom-up ‘design’ where there is no architect/choreographer, only low-level local rules.
Later stages of embryology are too complicated to deal with here. Different tissues – muscle, bone, nerve, skin, liver, kidney – all grow by cell division. The cells of each tissue look very different from each other, but all have the same DNA. The reason they are different is that different stretches of DNA – different genes – are turned on. In any one tissue, only a small minority of the tens of thousands of genes are turned on. What this means is that in each tissue, the proteins, those vital ‘lab-assistant’ enzymes, that are made in the cells of that tissue are only a small minority of the enzymes that could be made – and actually are made in other tissues. And that leads to the cells in different tissues growing differently. Each tissue grows by cell division following local bottom-up rules. And each tissue stops growing when it reaches the right size: again, following bottom-up rules. Sometimes things go wrong and a tissue fails to stop growing: cells disobey the bottom-up rules that tell them to stop dividing. That’s when we get a tumour, like a cancer. But mostly that doesn’t happen.
Now let’s put the idea of bottom-up embryology together with the crystals of Chapter 9. Crystals – pyrites or diamonds or snowflakes – grow their pretty shapes by local bottom-up rules. In those cases the rules are the rules of chemical bonds. We likened the molecules organized by those rules to soldiers on parade. The important point is that nobody designed the shape of the crystal. The shape emerged through the obeying of local rules.
Then we saw how the laws of chemical bonds – by a process that resembles jigsaw pieces slotting into each other – produced more elaborate things than ordinary crystals: protein molecules. Then the same kind of jigsawing caused the protein chains to coil up into ‘knots’. And the ‘crevices’ in the ‘knots’ enabled them to act as enzymes, catalysts that turn on very particular chemical reactions inside cells. As I said before, ‘crevices’ is a great oversimplification. Some of these knotted molecules are tiny machines, miniature ‘pumps’, or tiny ‘walkers’ which literally stride about on two legs inside the cell, busily doing chemical errands! Look on YouTube for ‘Your body’s molecular machines’ and be utterly amazed.
Enzymes switch on other enzymes which, in turn, catalyse other particular chemical reactions. And those chemical reactions inside cells cause the cells to work together, following local rules as in George Oster’s simulation, to make an embryo. And then a baby. And every step of the way is controlled by DNA, again using just the same jigsaw rules. It’s like crystals all the way through, but elaborate crystals of a very special kind.
The process doesn’t stop with birth. It goes on as the baby grows into a child, the child grows into an adult, and the adult grows older. And of course, differences in DNA in different individuals – ultimately caused by random mutations – cause differences in the proteins that ‘crystallize’ or ‘tie knots’ under the influence of the DNA. And the knock-on effects of those differences eventually show themselves, way down the line, in differences in the adult body. Perhaps the adult cheetah runs just a little bit faster. Or slower. Perhaps the chameleon’s tongue shoots out just that little bit further. Perhaps the camel can cover just a few more miles of desert before dying of thirst. Perhaps the rose thorn is just a tiny bit sharper. Perhaps the cobra’s venom is just a tad stronger. Any mutation in the DNA can have an effect, at the end of the long, long chain of intermediate effects on protein and cell chemistry and embryonic growth patterns. And that can make the animal more, or less, likely to survive. And that makes it more, or less, likely to reproduce. And that makes the DNA responsible for the change more, or less, likely to find itself in the next generation. So, as the generations go by, over thousands and millions of years, the genes that survive in the population are the ‘good’ genes. Good at building bodies that run fast. Or have long tongues. Or can go more miles without water.
That, in a nutshell, is Darwinian natural selection, the very reason why all animals and plants are so good at what they do. The details of what they are good at are different for each species. But it’s all ultimately about being good at one thing: surviving long enough to pass on the DNA that makes them good at whatever it is they do. After thousands of generations of this natural selection, we notice (or we would if we lived long enough) that the average form of the animals in the population has changed. Evolution has occurred. After hundreds of millions of years, so much evolution has happened that an ancestor looking like a fish has given rise to a descendant looking like a shrew. And after billions of years, so much evolution has occurred that an ancestor like a bacterium has given rise to a descendant like you or me.
Everything about a living creature is the way it is because its ancestors evolved that way over many generations. That includes humans and it includes human brains. The tendency to be religious is a property of human brains, as is the tendency to like music and sex. It’s therefore reasonable to guess that the tendency towards religious belief has an evolutionary explanation, like everything else about us. And the same goes for our tendencies, such as they are, to be moral, or to be nice. What might the evolutionary explanation be? That is the topic of the next chapter.
Until pretty recently just about everybody believed in some sort of god. Outside western Europe, where only a minority nowadays are religious, most people around the world, including the United States, still do believe in a god or gods, especially if they aren’t well educated in science. Shouldn’t there be a Darwinian explanation of belief in gods? Did religious belief, belief in some kind of god or gods, help our ancestors to survive and pass on genes for religious belief?
I suspect that the answer is probably yes. Well, a kind of yes. Of course that doesn’t mean that the gods people believe in – whichever gods those might be – are really there. That’s a completely separate question. Believing in something that isn’t really there could even save your life. There are various ways in which this might happen.
You remember the gazelles and zebras needing to strike a fine balance between being too scared and not scared enough? Now imagine you are an early human, long ago in our ancestral past on the African plains. Like a gaz
elle, you have to get the balance right between being sufficiently scared of lions and leopards, and being so scared that you never get on with the business of life. In the human case that might be the business of digging for yams or courting a mate. You hear a noise and look up from digging up a yam. You see a movement in the grass which just might be a lion. It could instead be the wind. You are making good progress in digging out a really big tuber and don’t want to stop. But that noise just could be a lion.
If you believe it’s a lion and it really is a lion, that valid belief might save your life. That’s easy to understand. The next part is harder to understand. Even if it is not a lion on this particular occasion, a general policy of believing that mysterious movements or sounds spell danger could save your life. Because sometimes it really will be a lion. If you take that too far and run scared from every rustle in the grass, you’ll miss out on the yams and the other business of living. But an individual who gets the balance right will still, on some occasions, find himself believing it’s a lion when it actually isn’t. And that tendency to believe what may turn out to be a falsehood will sometimes save your life. That’s one way in which believing in things that don’t exist could save your life.
Outgrowing God Page 16