Tuesday 18 December 2012

The Nose Cone versus the Tree

Years ago, when I was on a mini lecture tour of Californian universities, my host at Stanford took me out to dinner, and we got talking about his wife’s work as a space scientist.  She would compete to get room in the payload of the next space rocket launch, trying to win a small space allocation for her experiments. She would then have to pack as much instrumentation as possible into a small size and weight allowance. This involved juggling many parameters so as to maximise the scientific outcomes. All the other winning scientists would be doing the same with their own independent stand-alone experiments. Collaboration was pretty restricted. In the end the rocket blasted into space with a compact collection of scientific instruments. This is a good analogy for the theory that human brains are made from separate components, refined earlier in evolution, and packed into the human skull.

Another approach to space experimentation is to have a shared general processor which reduces duplication. The Pioneer 10 spacecraft used this sort of approach, being built as one coordinated system though carrying individual sensors. It remains our one only and very successful export from the solar system. Basic processes were done in common, and only specialised instruments require separate modules. Conceptually, coordinated structures look like a tree: a large trunk of common functions, two or three big branches of semi-specialised functions, and then little twigs of highly specialised instruments.

Human intelligence is more like a tree than a nose cone. We know this because if you pack separate modules in the confined space then each highly developed and larger module leaves less space for other components: aptitudes will be negatively correlated. A strong visual ability might compel language ability to be less developed, through lack of space, and so on. Each brilliant skill would lead to a scarcity of competence in other intellectual areas.

On the contrary, human intellectual skills always show a positive manifold, a matrix of positive correlations, such that there is a common factor which accounts for 50% of the variance. People who are good at one intellectual task tend to be above average at all other intellectual tasks.  It would appear that Mother Nature has got round the limitations of the Nose cone problem (skulls cannot get infinitely large) and has gone for a common central processor.

In humans this common factor is virtually always found, and by convention is referred to by a lowercase g. This is an abstraction, but represents the factor that all intellectual tasks have in common. Despite the accumulation of evidence, some still argue for a modular approach to skills, and think of g as an abstraction created by intelligence tests. So, can one show that g exists in other species?

Looking back at the records of 60 rhesus macaques on learning, spatial memory, object memory and set shifting, Rosalind Arden (Middlesex University) found that one factor accounts for 47% of their success at these mental tasks. In headline terms, this is a case of Monkey IQ.  Curiously, success was also associated with total cholesterol, an association which has also been found in humans.

Looking at the success of laboratory mice across 10 different tests of learning, reasoning, and attention, Louis Matzel (Rutgers University) found that a common factor accounted for 30-50% of the variance, a case of Mouse IQ. A large percentage of Mouse g was accounted for by a small number of dopaminergic genes in the medial and dorsolateral prefrontal cortex. Mouse mental skills were improved by working memory training.

It would appear that there is an evolutionary advantage to having a central processor in the skull, which can bear the brunt of problem solving. This frees up space for other essential functions like bodily coordination, visual processing and some more specialised functions. Given that there are physical problems about making brains bigger (every neurone requires connections, and the whole array is very energy consuming) it is a relief that Nature has used its intelligence.



  1. What would be compelling to some of us would be to demonstrate pig IQ.

  2. Aha, someone disagrees with you.

  3. Dear dearime, If you read the item carefully, particularly the last paragraph, you will see that Steve agrees with the reality of general IQ estimates, and is ironic about multiple intelligences, strongly suggesting that the two examples given were not of intelligently lived lives.

  4. Not Steve, the chump "scholar" is the one who disagrees with you. I'll have to assume you're suffering from jet lag.

  5. The Indy apparently Has Views on the subject of IQ.