seems mainly to consist of a difficulty in empathising with other people’s thoughts. Since boys are less good at empathising than girls anyway, then perhaps autism is just an extreme version of the male brain. Hence Baron-Cohen’s interest in the inverse correlation between prenatal testosterone and eye contact: the masculinisation of the brain by testosterone may go ‘too far’ in autistics.36
Intriguingly, Asperger’s children are often better than normal at folk physics. Not only are they frequently fascinated by mechanical things, from light switches to aeroplanes, but they generally take an engineering approach to the world, trying to understand the rules by which things – and people – operate. They frequently become precociously expert in factual knowledge and mathematics. They are also more than twice as likely to have fathers and grandfathers who worked in engineering. On a standard test of autistic tendencies, scientists generally score higher than non-scientists and physicists and engineers score higher than biologists. Baron-Cohen says of one brilliant mathematician, a winner of the Fields medal, who has Asperger’s: ‘Empathy passes him by.’37
To demonstrate how a difficulty with folk psychology can coexist happily with expertise at folk physics, psychologists designed two remarkably similar tests called the false-belief test and the false-photo test. In the false-belief test, the child sees the experimenter move a concealed object from one receptacle to another while a third person is not watching. The child then has to say where the third person will look for the object. To get the right answer, he has to understand that the third person holds a false belief. All children pass this test for the first time around the age of four (boys later than girls), but autistics are especially late developers.
In the false-photo test, by contrast, the child takes a Polaroid photograph of a scene, then, while the picture is developing, sees the experimenter move one of the objects in the scene. The child is asked which position the object will occupy in the photograph. Autistics have no difficulty with this test, because their understanding of folk physics outstrips their understanding of folk psychology.
Folk physics is just part of a skill that Baron-Cohen calls ‘systemising’. It is the ability to analyse input-output relationships in the natural, technical, abstract and even human world: to understand cause and effect, regularity and rules. He believes that human beings have two separate mental abilities, systemising and empathising, and that, though some people are good at both, others are good at one and bad at the other. Those who are good systemisers and bad empathisers will try to use their systemising skills to solve social problems. For instance, one person with Asperger’s said to Baron-Cohen that ‘Where do you live?’ was not a good question, since it could be answered on many levels: country, city, district, street or house number. True, but most people solve the problem by empathising with the questioner. If speaking to a neighbour, he might name the house; if to a foreigner, the country.
If Asperger’s people are good systemisers and bad empathisers, with extreme-male brains, the thought arises that there are probably people who are good empathisers and poor systemisers, with extreme female brains. A moment’s thought will confirm that we all know such people, but their particular skill combination is rarely classified as pathological. It is probably easier to live a normal life in the modern world with poor systemising skills than with poor empathising skills. In the Stone Age, it might have been less easy.38
A MIND IN PARTS
The empathy story illustrates a very William James theme of separate instincts. To be good at empathising you need a domain, or module, in your mind that learns to intuitively treat animate creatures as having mental states as well as physical properties. To be good at systemising, you need a domain that learns how to intuit cause and effect, regularities and rules. These are separate mental modules, separate skills and separate learning tasks.
The empathy domain seems to rely on circuits around the paracingulate sulcus, a valley of the brain close to the mid-line and near the front of the head. In the studies of Chris and Uta Frith in London, this area lights up (in a suitable scanner) when a person reads a story that requires ‘mentalising’ – imagining the mental states of others; it does not light up when the person reads a story about physical cause and effect or a series of unlinked sentences. In people with Asperger’s syndrome, however, this area does not light up when reading mental-state stories, but a neighbouring area does, instead. This is an area implicated in general reasoning, which supports the psychologists’ hunch that Asperger’s people reason about social issues rather than empathise about them.39
All of which rather supports the idea that Jamesian instincts must be manifest in mental circuits called modules, each specifically designed to be good at its specific mental task. Such a modular view of the mind was first enunciated by the philosopher Jerry Fodor in the early 1980s and later developed by the anthropologist John Tooby and the psychologist Leda Cosmides in the 1990s. Tooby and Cosmides were attacking the then widespread belief that the brain is a general-purpose learning device. Instead, said the anthropologist-psychologist couple, the mind is like a Swiss army knife. For blades and screwdrivers and things for helping Boy Scouts get stones out of horses’ hoofs, read vision modules, language modules and empathy modules. Like the tools of a knife, these modules are rich in teleological purpose: it makes sense not just to describe what they are made of and how they do their job, but what they are for. Just as the stomach is for digestion, so the visual system of the brain is for seeing. Both are functional, and functional design implies evolution by natural selection, which implies at least partly a genetic ontology. The mind therefore consists of a collection of content-specific, information-processing modules adapted to past environments. Nativism was back.40
This was the high point of what is sometimes called the cognitive revolution. Though it now owes much to the tragic genius Alan Turing, with his extraordinary mathematical proof that reasoning could take a mechanical form – that it was a form of computation – the cognitive revolution really began with Noam Chomsky in the 1950s. Chomsky argued that the universal features of human language, invariant throughout the world, plus the logical impossibility of a child deducing the rules of a language as quickly as it does merely from the scanty examples available to it, must imply that there was something innate about language. Much later, Steven Pinker dissected the human ‘language instinct’, showed it had all the hallmarks of a Swiss army knife blade – structure designed for function – and added the notion that what the mind was equipped with was not innate data, but innate ways of processing data.41
Do not mistake this for an empty or obvious claim. It would be quite possible to imagine that vision, language and empathy are done by different parts of the brain in different people. That indeed is the logical prediction that follows from the empiricist argument that runs from Locke, Hume and Mill right up to the modern ‘connectionists’ who design multi-purpose computer networks to mimic brains. And it is wrong. Neurologists can produce battalions of case histories to support the idea that particular parts of the mind correspond to particular parts of the brain with very little variation all over the world. If you damage one part of your brain, in an accident or after a stroke, you do not suffer some generalised debility: you lose one particular feature of your mind – and the feature you lose depends precisely on which part of the brain is lost. This cannot but imply that different parts of the brain are pre-designed for different jobs, something that could only come about through genes. Genes are often thought of as constraints on the adaptability of human behaviour. The reverse is true. They do not constrain; they enable.
True, there have been rearguard actions by the retreating empiricists, but these skirmishes have only briefly delayed the advance of the modular mind. There is a degree of plasticity in the brain
