Science of Discworld

ANTHILL INSIDE

You KNOW WHAT'S GOING TO HAPPEN TO THE APES -they're going to turn into us. But why do we have them playing in the surf? Because it's fun? Yes ... but more significantly, because the seashore is central to one of the two main theories about how our ape ancestors acquired big brains. The other, more orthodox theory places the evolution of the big brain out on the African savannahs, and we know that some of our ancestors lived on the savannahs because we've found fossils. Unfortunately, seashores aren't a good place to leave fossils. You often find them there, but that's because they were deposited when the area wasn't a seashore at all, and the sea has subsequently eroded the rocks to expose the fossils. In the absence of direct evidence of this kind, the surfing apes theory has to take second place ... but it does explain our brains rather neatly, whereas the savannah theory rather sidesteps this issue.

Our closest living relatives are two species of chimpanzee: the stan­dard boisterous 'zoo' chimp Pan troglodytes and its more slender cousin the bonobo (or pygmy) chimp Pan paniscus. Bonobos live in very inaccessible parts of Zaire, and weren't recognized as a sepa­rate species of chimpanzee until 1929. We can to some extent unravel the past evolutionary history of the great apes by compar­ing their DNA sequences. Human DNA differs from the DNA of either chimpanzee by a mere 1.6%, that is, we have 98.4% of our DNA sequences in common with theirs. (It is interesting to specu­late on what the Victorians would have made of this.) The two species of chimpanzee have DNA that differs by only 0.7%. Gorillas differ from us, and from both chimps, by 2.3%. For orang­utans, the difference from us is 3.6%.

These differences may seem small, but you can pack an awful lot into a small percentage of an ape genome. A big chunk of what we have in common must surely consist of 'subroutines' that organize basic features of vertebrate and mammalian architecture, tell us how to be an ape, and tell us how to deal with things we've all got -like hair, fingers, internal organs, blood ... The mistake is to imag­ine that everything that makes us human and not a chimpanzee must live in that other 1.6% of 'special' DNA, but DNA doesn't work that way. For example, some of the genes in that 1.6% of the genome may organize the other 98.4% in a completely new way. If you look at the computer code for a wordprocessor and a spread­sheet, you'll find they have an awful lot in common, routines for reading the keyboard, printing to the screen, searching for a given text string, changing fonts to italic, responding to a click on the mouse ... but this doesn't mean that the only distinction between a spreadsheet and a wordprocessor lies in the relatively few routines that are different.

Since evolution involves changes to DNA, we can use the sizes of those differences to estimate when various ape species diverged from each other. This method was introduced by Charles Sibley and Jon Ahlquist in 1973, and while it needs to be interpreted with caution, it works well here.

A convenient unit of time for such discussions is the 'Grandfather', which we define to be 50 years. It's a good human length, being about the age difference between the child and the grandparent who says 'When I was young ...' and passes on a sense of history. In these terms, Christ lived 40 Grandfathers ago, and the Babylonians go back about 100 Grandfathers. That's not a lot of grandads, passing down through recorded human history recollec­tions like '... we never had any of this modern cuneiform when I was a lad ...' and'... bronze was good enough for me'. Human time is not very deep. We've just been good at packing a lot into it.

DNA studies indicate that the two chimp species diverged about 60,000 Grandfathers ago, three million years. Humans and chimps diverged 80,000 Grandfathers earlier, so a chain of only 140,000 grandfathers unites you and your chimplike ancestor. Who was also, we hasten to point out, a modern chimpanzee's manlike ancestor. Humans and gorillas diverged 200,000 Grandfathers ago; humans and orangutans diverged 300,000 Grandfathers ago. So among these animals, we are most closely related to a chimpanzee, and least closely related to the orangutan. This conclusion is borne out by physical appearance and habits, too. Bonobos really like sex.

If those times seem rather short for all the necessary evolution­ary changes, bear two things in mind. First, that they were estimated by using a realistic rate for DNA mutations; second, that according to Nilsson and Pelger an entire eye can evolve in a mere 8,000 Grandfathers, and lots of different changes can, should, and did evolve in parallel.

The most striking feature of humans is the size of our brains: bigger, in comparison to body weight, than any other animal. Strikingly bigger. A detailed story of what makes us human must be extraordinarily complicated, but it's clear that big, powerful brains were the main invention that made it all possible. So we now have two obvious questions to think about: 'Why did we evolve big brains?' and 'How did we evolve big brains?'

The standard theory addresses the 'why'. It maintains that we evolved out on the savannahs, surrounded by lots of big predators, lions, leopards, hyenas, and without much cover. We had to become smart in order to survive. Rincewind would instantly see one flaw in this theory: 'If we were so smart, why did we stay on the savannahs, surrounded by lots of big predators?' But, as we've said, it fits the fossil evidence. The unorthodox theory addresses the 'how'. Big brains need lots of brain cells, and brain cells need lots of chemicals known as 'essential fatty acids'. We have to get these from our food, we can't build them ourselves from anything sim­pler, and they're in short supply out on the savannahs. However, as Michael Crawford and David Marsh pointed out in 1991, they are abundant in seafood.

Nine years earlier Elaine Morgan had developed Alister Hardy's theory of the 'aquatic ape': we evolved not on the savannahs, but on the seashore. The theory fits a number of human peculiarities: we like water (newborn babies can swim), we have a funny pattern of hair on our bodies, and we walk upright. Go to any Mediterranean resort and you see at once that an awful lot of naked apes think that the seashore is the place to hang out.

Brains are fascinating. They are the physical vehicle for minds, which are even more interesting. Minds are (or, at least, give their owners the vivid impression that they are) conscious, and they have (or, at least, give their owners the vivid impression that they have) free will Minds operate in a world of 'qualia', vivid sense impres­sions like red, hot, sexy. Qualia aren't abstractions: they are 'feelings'. We all know what it's like to experience them. Science has no idea what makes them the way they are.

Brains, though ... we can make progress on brains. On one level, brains are a kind of computational device. Their most obvious physical components are nerve cells, arranged in complicated net­works. Mathematicians have studied such networks, and they find that what networks do is to carry out interesting processes. Give them an input and they will produce an output. Allow their inter­connections to evolve by selecting for specific associations of input and output, such as responding to an image of a banana but not to an image of a dead rat, and pretty soon you've got a really effective banana-detector.

What makes the human brain unique, as far as we can tell, is that it has become recursive. As well as detecting a banana, it can think about detecting a banana. It can think thoughts about its own thought processes. It is a pattern-recognition device that has turned its attention to its own patterns. This ability is what lies behind human intelligence. It probably underpins consciousness, too: one of the patterns that the pattern-recognition device has learned to recognize is itself. It has become 'self-aware'.

As a result, brains operate on at least two levels. On a reduc­tionist level they are networks of nerve cells sending each other incredibly complex but ultimately meaningless messages, like ants scurrying around inside an anthill. On another level, they are an integrated self, the anthill as a personality in its own right. Douglas Hofstadter's Godel, Esther, Bach includes a sequence where Aunt Hillary (who is an anthill, use the American pronun­ciation of 'Aunt') has a meeting with Dr Anteater. When Dr Anteater arrives, the ants go into a panic, they change their actions. To Aunt Hillary, who operates on the emergent level, this change represents the knowledge that Dr Anteater has arrived. She is entirely happy to watch Dr Anteater consuming a meal of 'her' ants. Ants are a virtually inexhaustible resource, she can always breed new ones to take the place of the ones that got eaten.

The link between the ants and Hillary's 'anthilligence' is emer­gent, felicitously, it. operates across what we have termed 'Ant Country'. The same action means one thing for the ants, but some­thing quite different, and transcendent, for Hillary. Replace Hillary by yourself, your self, the 'you' that you feel is experiencing your thoughts, and ants by brain cells, and you're contemplating the connection between mind and brain.

Now you've gone self-referential.

Neural networks are what the brain is built from, but there's more to evolving a brain than just assembling big neural nets. Brains operate in terms of high-level 'modules', a module for running, another for recognizing danger, another for putting the whole ani­mal on the alert, and so on. Each such module is an emergent feature of a complex neural network, and it wasn't designed: it evolved. Millions of years of evolution trained those modules to respond instantly and exquisitely.

The modules aren't separate. They share nerve cells, they over­lap, they're not necessarily a well-defined region in the brain, any more than 'Vodafone' is a well-defined region of the telephone net­work. According to Daniel Dennett, they are like a collection of demons, operating by 'pandemonium'. They all shout, and at any given instant, whoever shouts loudest wins (quite a lot of the Internet has borrowed this design).

Modern humanity has built a culture around those modules, an idea that we'll explore later, and in so doing has subverted them to new purposes. The module for spotting lions has become, in part, a module for reading Discworld books. The module for sensing bodily movement has, in part, turned into one for doing certain kinds of mathematics, those parts of mechanics where a physical 'feel' for the problem may well be precisely that. Our culture has rebuilt our minds, and our minds have in turn rebuilt our culture, over and over again, in each generation.

Such a radical restructuring must have simpler precursors. A key step towards the human mind was the invention of the nest. Before there were nests, baby organisms could carry out only very limited experiments in behaviour. If every time you try out a new game you get gobbled up by a python, novelty will not carry a pre­mium. In the comfort and relative safety of the nest, however, the error part of trial-and-error is no longer automatically fatal. Nests let you play, and play lets you explore the phase space of possible behaviours and find new, sometimes useful, strategies. Further along the same path lies the family, the pack, and the tribe, with cer­tain shared behaviours and mutual protection. Meerkats, a kind of mongoose, have an intricate tribal structure, and take turns doing the dangerous (because more exposed) job of Lookout.

Humans have turned such tactics into a global strategy: adults devote huge amounts of time, energy, food, and money to the task of bringing up their children. Intelligence is both a consequence of this brilliantly successful strategy, and a cause.

The Dean would be well advised to take this link between family life and intelligence into account. He's trying to educate the apes by the direct route (R ... O ... C ... K ...) but all they have on their tiny minds is S-E-X. Many school teachers will sympathize ... but if only he realized that sexual bonding is a major factor in humanoid family life, and family life engenders intelligence ,..

Bonobos are the perfect model for the Dean's sex-mad apes. They are promiscuous in the extreme, making use of sex where we would be content with a smile and a wave or a gentlemanly hand­shake. Female bonobos have serial sex with dozens of males, or with females, almost in passing; the males do likewise. Adults engage in sexual activities with children, too. It all seems very casual. It helps bond the tribe. For them it seems to work fine.

Ordinary chimps are promiscuous by the standards of orthodox human morality, though probably no more than many humans are. Pairs of males and females will disappear together for a few days, and then form new partnerships ... Humans generally mate for life (a term meaning 'until we get fed up') and one reason is the enor­mous amount of effort that a human couple must put into raising the kids. Sex helps to cement the parental relationship, encouraging each parent to trust the other. This .may be why, even in an allegedly sexually relaxed age, most people see extramarital flings as a form of betrayal, and why, despite that, the erring partner is more often than not allowed back into the family fold.

It's not surprising that we have sex on the brain: our brains have been moulded by sex. The Dean should let sex take its course, for intelligence will surely follow ... You just have to think on the scale of Deep Time. There's no rush.