For most of the last century it was widely thought that intelligence was in decline.
The idea was that those at the lower end of the intelligence spectrum were having more children, thereby reducing the general intelligence level.
Then, one November day in 1984, a New Zealand-based moral philosopher named James Flynn had a eureka moment that turned cognitive science on its head. He opened a package sent to him by an academic in Holland named P.A. Vroon.
Inside was data from an IQ test known as Raven's Progressive Matrices. IQ stands for "intelligence quotient", the psychometric system by which the mental ability known as intelligence is measured. Vroon did not know how to crunch the raw numbers, but Flynn did.
And he noticed that 18-year-old Dutch males had made a giant leap in IQ scores on the previous generation. Over the following month, he checked similar data from around the world and the answer was the same: IQ was going up, dramatically.
The American-born Flynn found that in both the developed and developing worlds IQ had improved in the 20th century at the remarkable rate of 3 points per decade. This development has since become known as the "Flynn effect".
It is noteworthy because, apart from anything else, it suggests a world-changing increase in intelligence in succeeding generations. IQ measurements are based on the score of 100 being allocated to the median average of a group (for example, 18-year-old Dutch males). Projecting forwards, the Flynn effect predicts, for example, that someone with an average IQ today (of 100) will have grandchildren with a score of 120.
Perhaps more shocking, it suggests that someone with an average score today would have had grandparents who were close to mental retards. Neither scenario makes much sense, particularly if you're a grandchild or a grandparent.
As Flynn writes in his recent book, What is Intelligence?: "In either event, the cognitive gulf between the generations should be huge." Plainly it's not.
And thus the Flynn effect has become one of those phenomena that are almost universally accepted but little understood. Some psychologists have attempted to explain the gains in what Flynn calls "artefactual" terms. It has been argued that children have got better at doing the tests partly because the administration of the tests has improved.
Others put the rise in IQ down to improvements in diet. Flynn rejects both theories, pointing out that tests are frequently restandardised and that diet has been of marginal importance in the West since the 1950s. Flynn - an emeritus professor of political studies at the University of Otago - is convinced that intelligence has improved.
"But," he says, "we have to rethink exactly what we mean by intelligence. What the IQ gains really give us is a cultural history of the 20th century and an insight into the gulf that separates our minds from those of our ancestors."
He uses the phrase "detaching logic from the concrete" to define what is considered the mark of intelligence in the modern world. "Imagine," says Flynn, "if we were archaeologists of the future and all we found were targets from target practice.
We would see that in 1880 there were two shots per minute. In 1900 there were five shots. And later in the century, 50 shots. How could they have improved their aim, speed and steadiness by such an amount? We'd be totally baffled until we dug up pistols, rifles and machine guns. So it's not a change in human physiology.
The same applies for intelligence. We've created tools and the environment to maximise those scores." In What is Intelligence? he gives a number of examples of how we now classify abstract ideas.
A century ago if the question was asked what links dogs and rabbits, the most common answer would probably have been that dogs are used to catch rabbits.
That's a zero-point answer in an IQ test, where the correct answer is that both animals are mammals. But if that seems a cultural rather than a cerebral development, the change is not just cosmetic. The ability to think in abstract terms underpins the scientific, material and, arguably, moral advances of modern society.
Therefore the continued rise in abstract intelligence might be expected to produce a continued rate of scientific and material development. Except there are signs that the Flynn effect is coming to an end, at least in western Europe.
Recent studies in Scandinavia show a plateau in results and a drop in arithmetic. Far from being surprised, Flynn has been expecting as much. "[The Flynn effect] is not like the law of gravity," he says. "Social conditions vary in each country."
One of the key environmental factors, Flynn says, is adult-infant ratio. Basically, the more adults there are to children, the richer the cognitive environment.
As you can't really get a better ratio than the two parents to one child, there is not much further the low birth rate in Europe can go in that direction. Indeed, the rise of solo parents may push the trend in the opposite direction.
Flynn also believes that Scandinavian teaching methods are the best around, and perhaps they, too, have gone as far as they can go. "There's got to be a limit to how much we can classify the world and to how cognitively challenging our leisure time can be. Everyone needs to relax sometime.
And if IQ levels in the West slow down and those in the developing world continue to increase and catch up, then we could be in for some interesting times."
Another study by British psychologist Michael Shayer, of King's College, University of London, looked at tests concerned with volume and heaviness.
It showed a marked reversal in geometric reasoning. In 2003 children of almost 12 years did as well as 8- or 9-year-olds in 1976. The biggest drop was in the performance of boys.
Shayer believes that boys today are less inclined to develop the "differential play patterns" that previously accounted for their advantage over girls. In short, they have grown less prepared to explore further afield, to go beyond the comfort zone of their controlled environments.
"Presumably," says Shayer, "because they were looking at bloody computer games." Flynn thinks computers can help with abstract cognitive skills, but, he warns, there is a price.
"They don't read, the little bastards," he says of young people today. "And I don't consider someone educated unless they can read Tolstoy or Plato." For his part, Shayer is sceptical about the value of the psychometric tests by which IQ is measured.
"All they will tell you is how someone compares with a representative group of that age," he says. "They won't tell you what they can do."
What constitutes intelligence is a question over which generations of psychologists and educationalists have argued. And there is still no conclusive answer.
All we can state with confidence is what is not intelligent. For example, it's not intelligent to stay up drinking alcohol until the early hours the night before sitting an IQ test. That, as it turned out, was my second mistake in preparation for the Mensa test. My first mistake was to not practise.
Instead, I walked into the examination cold. It was the first time in more than 20 years that I had put myself through any kind of formal mental test. I felt anxious.
Scaled intelligence testing was first introduced by the French psychologist Alfred Binet in 1905, though it was his German colleague William Stern who coined the abbreviation IQ. The number of people sitting the Mensa test each year is rising as various forms of psychometric testing become more commonplace in business and employment.
I was not sure that I would find out what my IQ measured by sitting the test. But I did hope that it would at least help me get a better grip on the question.
And, okay, I won't deny it, I've always been secretly curious to know my IQ rating. There are two papers in the Mensa test. The first is called the Culture Fair test, which assesses non-verbal reasoning through a series of diagrammatic problems.
The other is called the Cattell B test, which is more concerned with verbal reasoning, though it also features a number of visual-only questions as well. To be invited to join Mensa, you must score in the top two per cent of the population in either one of the tests. For the Culture Fair test this means a score of 132 or above and in Cattell B at least 148.
Other IQ tests have different rating systems. So when you hear talk of people having a certain IQ rating, it doesn't tell you much, unless you know which test they took. Mensa was founded in Britain in 1946 by Australian barrister Roland Berrill and British lawyer-cum-scientist Dr Lancelot Ware.
There are now around 100,000 members around the world, of which a quarter are in Britain and half in America. It's a non-political organisation whose stated purpose is to encourage research into intelligence and create a social forum for its members.
"The word Mensa is not an acronym," explained the invigilator, before the test. "It's Latin for table, and it represents the round-table equality of Mensa. Everyone is equal. There is no elitism at Mensa." It seemed strange that an organisation that excludes 98 per cent of society should claim that it's not elitist.
But perhaps that's a prerequisite for any organisation these days. Mensa is, of course, specifically, constitutionally elitist, but it's also diverse. There were 15 people sitting the test: 11 males and four females from a variety of ethnic backgrounds.
They all looked clear-eyed and razor-sharp, the sort of people who could complete cryptic crosswords in a few minutes while reciting the decimal expansion of pi to a few thousand digits. The youngest was 12 and I, at 45, looked the oldest by a decade.
And therefore the slowest. Or so it seemed during the Culture Fair test. It started easily enough. Three simple shapes are presented and you have to select from six choices the correct shape that completes or fits into the pattern.
So, for example, the first shape is a black circle, the next is circle with three quarters black, and the north-west quarter white; the one after that, the east semi-circle black and the west white. And so the missing shape is a white circle with a black north-east corner.
But after that there was an exponential growth in geometric complexity until, a few questions on, I was looking at a series of diabolically intricate shapes that offered no obvious answer.
It was at this point, staring cluelessly at the abstract patterns, that I was overwhelmed by a profound sense of regret at the previous night's indulgence. Through the cognitive fog, there flickered an intuitive sense of the correct answers, but instead of ticking boxes I tried to work out proof to my satisfaction.
That took too much time and at the end of the half hour, I still had seven or eight questions that I hadn't answered. The invigilator had advised us to tick a random box if we didn't know an answer because a wrong selection doesn't count against you and there's a chance that you might pick the right box.
This seemed to introduce chance, therefore imprecision, into the assessment and, out of a mixture of high-mindedness and incompetence, I left blank the questions of which I was unsure.
I decided not to dilly-dally on the second test, the Cattell B, which is about an hour long. But halfway through, I felt the stirrings of a nervous breakdown. The intense level of concentration required was achingly beyond my experience. No getting up to make a cup of tea or surfing the internet, much less my preferred means of task avoidance: staring vacantly out of the window.
After a while, the questions began to take on different lives to those intended. "If Mr Jones is 40 years younger than his father and his father will be twice his age in 20 years, how long before ..." - I found myself wondering if Mr Jones visited his father at Christmas and whether Mr Jones Snr tended to repeat details and anecdotes.
The invigilator said it would be 10 days or so before we found out if we had attained the necessary score to join Mensa. I had no plans to hold my breath.
The phrase "Flynn effect" was first coined by Richard Herrnstein and Charles Murray in their 1994 bestselling book The Bell Curve. The book argued that IQ was a vital factor in a whole range of social issues, including employment, crime and single-parent families.
It also placed pronounced emphasis on the differences in average IQ scores between black and white Americans and suggested that these differences were in some sense innate. A number of scientists and psychologists have subsequently argued that much of the research underpinning their findings was flawed or skewed, pointing out that it drew on the research of eugenicists.
The controversy unleashed by The Bell Curve continues to dog the study of intelligence. Last year, James Watson, the Nobel prize-winning biologist who co-discovered DNA, triggered a whole new round of the dispute when he said that he was "gloomy" about Africa's prospects because "all our social policies are based on the fact that their intelligence is the same as ours - whereas all the testing says 'not really'."
Murray and Watson are what might be called genetic determinists. They look at the various IQ data relating to different ethnic groups and conclude that race is the cause of the variation. Watson's comment on Africa occupies the crude end of this argument.
There are obvious environmental reasons - poverty, lack of schooling, the less pressing importance of abstract reasoning in every day life - why Africa might lag behind in the IQ league that have nothing to do with gene distribution.
As Flynn says: "This idea that developing countries are too dumb to industrialise - well, our [relative] IQs were the same 100 years ago and we industrialised. It's like climbing a ladder. One foot goes up one rung - industrialisation - and that leads to the next rung - improving schools, etc."
It's worth remembering that the key findings behind the Flynn effect are not in dispute: IQ has been going up. But that does seem to call into question the very notion of innate, hereditary intelligence. For if IQ is simply handed down the generations, how is it that IQ scores rose so rapidly in the last century?
In his 1981 book The Mismeasure of Man, Stephen Jay Gould, the late zoology professor at Harvard, declared "the theory of unitary, innate, linearly rankable intelligence" to be full of "fallacies". The often-heard version of this view, much like Shayer's comment, is that psychometric intelligence testing only measures the ability to perform psychometric tests.
In any case, the steep rise in IQ would seem to lend weight to the idea that IQ does not evaluate some objective totality of intelligence but rather a cultural definition loaded towards the kind of abstract thinking that forms the basis of everyday life in post-industrial societies.
In which case a discrepancy between Africa and the West, or poor and rich (or its too-frequent manifestation, black and white) is effectively built in to the system. As the writer Malcolm Gladwell put it, IQ "measures not so much how smart we are as how modern we are".
So the question is, how modern am I? I certainly felt ancient during the tests and when my result arrived a week later I was reluctant to open it. Just under half those who take the Mensa test are invited to join (around 70 per cent take up the offer), but it's a self-selecting group. People tend to take the test because they think they'll make the grade.
The paradox was that before I took the test, I had no great expectations. But after it was finished, and I felt as if I'd fared poorly, I really wanted to have done well. I asked myself why I should care. After all, I didn't want to join Mensa.
It seemed like a perfectly benign and perhaps even worthwhile organisation, but I've always experienced a reflex wince whenever I've heard that someone - a celebrity or acquaintance - is a member of Mensa. It's a bit like those people who insist on telling you their excellent A-level results, 30 years after the event.
According to Flynn, when the philosopher Karl Popper was invited to join Mensa, "he said, whatever his intelligence was, it wasn't so low as to make him want to sit around with people whose only attribute was they the did well on IQ tests". Flynn agrees.
"I can understand why you might want to join up with people who share a love of music or literature. But to get together because you all did well in an IQ test? Bizarre." I asked some of the other participants why they wanted to find out their IQ.
Almost everyone cited curiosity as their reason. But one 14-year-old told me he thought it might help with job prospects. It was somehow depressing to see a boy preparing for the competitive job market before his voice had broken. I felt concerned that he appeared to have lost his corporate virginity at such a tender age.
What if he didn't qualify for Mensa, how would that make him feel? He gave me a look that was awkwardly close to contempt. "I'd feel really stupid," he said, "but that's not going to happen." Oh, for such confidence. "Your Supervised Test Result" was the headline of my Mensa result letter. My eyes shot down to the body copy and fixed on "113", my score in the Culture Fair test. That placed me in the 21st centile. Damn that hangover. No, let's be honest. It wasn't the hangover, it was me. Too slow to decode a few simple patterns.
So what? How often was I required to perform diagrammatic pattern-recognition tasks in real life? What good was that skill when parking a car or making coffee or even reading Tolstoy? Then I saw the "149" score of the Cattell B test, which put me in the second per centile. "We offer you our congratulations," continued the letter, "and would like to invite you to become a member of Mensa, the high IQ society." Yes! Phew, now I could turn them down, which was so much more appealing than the alternative: them turning me down.
Apparently journalists often do better on the Cattell B, because we regularly - at least in theory - use language-based reasoning. I found the two scores reassuring.
They seemed to demonstrate the subjective nature of IQ testing. If there's that level of variation in just one person, how can you draw definitive conclusions about whole groups?
The mysteries of the brain are not so easily reduced to a number. Naturally, if asked, I'll say I have an IQ of 149, rather than 113. It sounds better. But what does it amount to in the real world?
Something close, I suspect, to a great big zero.