g whiz! IQ is a clver concept

December 3, 1999

What is intelligence and is it all inherited? Such questions lead many academics to fear IQ research.In a week when researchers from around the world met in London to debate the issue,Rosalind Arden argues that we must face up to scientific truths

It can take a while to get through to people. It took me about a year to get Robert Plomin, a behavioural geneticist who works on intelligence at London's Institute of Psychiatry, to return my call. Calls, I should say. I work on science documentaries and had been gestating the idea of a film about intelligence for Channel Four's Equinox series and a slightly surreptitious baby.

Plomin's team identified a gene variant associated with high intelligence in 1996 and is trying to find other genes that contribute to intelligence. It is controversial research. His reluctance to talk to journalists stood me in good stead finally: it gave me time to read the literature and to bounce along with my rash but wonderful baby. Her sometimes shrieking presence also kept me aware of many of the issues that sussurate in the corridors outside the rooms where intelligence is researched.

The Novartis Foundation is host this week to a meeting on intelligence - a subject that until recently was viewed as distinctly off-colour. Speakers at the three-day discussion meeting, chaired by psychiatrist Sir Michael Rutter, include (my partner) Geoffrey Miller from University College London, Arthur Jensen, the US psychologist and key figure in intelligence research, and Douglas Detterman, who edits the journal Intelligence.

One hopes that fur and feathers will fly. Novartis is putting evolutionary psychologists in a room with behavioural geneticists. The two disciplines peer down opposite ends of life's binoculars, both focusing on human behaviour. Evolutionary psychologists think primarily about people at a species level - considering universal human characteristics such as language. Behavioural geneticists ponder differences between individuals, such as variation in language ability. The purpose of the meeting is to discuss why, if intelligence is a useful adaptation, there are still individual differences between people, given that millennia of evolution might have caused us to become alike.

It was precisely this study of why some people are more intelligent than others that made intelligence research the erstwhile ugly sister of the behavioural sciences. The siege mentality widespread in the field stems from experiencing years of hostility to biological - especially genetic - explanations of human variation. The long shadow of the Nazis is cast over almost all popular treatments of intelligence research. It is salutory to remember how vulnerable we can be to demonic ideologies; but Nazi thinking is not science now and it was not science then.

Preparing for the Equinox film taught me a lot about prejudices against IQ testing and studies of the intelligence of twins, because, to begin with, I shared all those negative views. I spotted quickly that there was a huge gap in understanding between most academics and the majority of psychologists, including most clinical practitioners working in settings such as hospitals or prisons. While the former considered the concept of IQ to be lifeless, the latter treated it as among the most vital instruments in their analytical toolkit.

My questions about intelligence research were fairly typical, I suspect. I wanted to know what exactly IQ tests were supposed to be measuring. And why, if so-and-so was so darn bright, did he appear to be such an imbecile? What did they mean by intelligence anyway? I was fortunate that the scientists I met did not try to persuade me of anything but they did send and recommend enough papers and books to make a houseful of draught excluders. So in the privacy of my own home, I learnt that a precise definition of intelligence is not the starting point. As with all science, it might be an end point, but it certainly is not the beginning.

A good starting point is the observation that in general our cognitive abilities overlap. The correlation is moderate, not perfect. As an example, you could take any teenager and ask him to perform a series of tests. These could range from vocabulary tests to driving tests. Though his performance on those tests would vary, there would be a moderate, positive correlation between the test scores.

The god of the overlap in mental abilities is the construct known as g. It is a construct because it arises from measurements, but that does not mean it is an artefact of those measurements. Schizophrenia is also a construct - it is a diagnosis aided by the use of various assessment tools, but it is real and appears to have a complex organic basis.

Named by Charles Spearman, g stands for "general factor" and is the core discovery of intelligence research. I have come to admire g enormously. It is so much less confusing than talking about intelligence - with all the baggage that entails. g is the active ingredient of intelligence. It is a bit like discovering the glucose molecule after struggling with comparisons of sweetness in countless cakes and colas. The construct g allows you to measure statistically the "g loading" of any cognitive task. This allows for proper comparisons between diverse activities so that playing bridge, shopping, taking care of pets or frowning over Raven's Progressive Matrices can be weighted based on their "g loading".

No one I have met thinks of g as biologically unitary or localised in a specific part of the brain. But it does have biological correlates including brain size, general health and cerebral glucose metabolism. It is likely that individual differences in g arise through a number of small contributing factors, such as the amount of myelination around neurons. Myelin is a fat that sheaths neurons, thereby aiding conductivity and preventing loss of signal strength. Britt Anderson from Alabama will be speaking at the Novartis meeting about g in non-human animals. As any pet owner knows, animals in the same species vary in responsiveness, learning ability and what most of us would call intelligence.

The heritability of intelligence is a stumbling block for many outside psychology. In the field there is a widely shared view that intelligence is highly heritable and that it is fairly stable throughout life. Heritability refers to the share of the differences between people that can be accounted for by their genes. Most human behavioural characteristics show some degree of heritability. Various researchers favour differing estimates of the heritability of intelligence, but the meta-studies show a figure of about 50 per cent. The crunch finding, though, is not that genes account for half of the differences between individuals in intelligence. It is that we have no idea what on earth causes the remainder of the difference.

We have all read enough articles about gene-versus-environment dichotomies. But what exactly is the "environment"? It has a simple but technical meaning - "non-genetic factors". Those include biological variables such as the uterine environment, nutrition, effects of alcohol and smoking on the foetus, plus the more usual influence of parents and socioeconomic conditions. What gets almost no publicity is the fact that, despite decades of intensive research, no one has been able to identify a single factor in the external, manipulable environment that makes a damn bit of difference to the g for any person who lives within the normal range of environments. That is not true for people whose lives are coloured by neglect and abuse.

Why are we so afraid of intelligence research? We do not want our children to be "labelled", we do not want scientists to tell us about our child's intellectual trajectory from birth even if they could (which they cannot). We do not want our school funds taken away from the needy and focused on a genetic ubermensch. I have come to view intelligence research as less the ugly sister and more the Prince Charming. Our policies should reflect our values, but to achieve our goals, we are better equipped by having a clear description of "what is" in order to arrive at "what we choose". Most of us would like to see all children having access to a good education with equal access to future job opportunities. How do we achieve that? We are not equal in size (which was more important in our ancestral past) and we are not equal in abilities. I have never met a parent who thought otherwise, yet we act as if by pretending we were, somehow we serve our children better.

By failing to recognise individual differences in intelligence and their causes, I think we let children down. At the moment we have a one-size-fits-all approach to schooling. We have arranged all our classes on an age basis rather than an ability basis, so those children who deviate from the age norm in ability are badly served. The proven validity of standardised testing for intelligence could improve the fit between teaching level and child. Indeed, Southhampton University's Stevan Harnad, one of the Novartis discussants, has suggested that supplementing A levels as a qualification for university with US academic achievement and aptitude tests would be an improvement. A levels have the merit that students can score well by working hard, given some ability. Standardised testing might benefit students who have had poor schooling but who are innately bright and could make good use of a well-stocked university library.

Does anything we know from current behavioural genetics have implications for today? Intelligence research has moved from the Salon des Refuses to Portland Place. Some of the discussants are former opponents of g. Today they are won over by evidence that g exists, is measurable and partly heritable. As London School of Economics academic Nicholas Humphrey says:

"We have to learn to accept the facts and understand the implications." It is nearly the new century, it is time to use the information we have towards creating the society we wish to live in.

Rosalind Arden and Robert Plomin are writing a book on Intelligence and g, to be published by Weidenfeld and Nicolson in spring 2001.

Register to continue

Why register?

  • Registration is free and only takes a moment
  • Once registered, you can read 3 articles a month
  • Sign up for our newsletter
Please Login or Register to read this article.