This recent study involving the testing of intelligence in autistic individuals raises many interesting and controversial concepts worthy of discussion – although I don’t anticipate arriving at definitive resolutions. The study involves comparing a standard IQ test, which is heavily dependent upon language function, to a novel test of intelligence that involves “inferring abstract rules and thinking abstractly about geometric patterns.” In the former case, many autistics score below average, even at or below the cutoff for “low functioning.” In the latter case the same individuals may score at or above average. What does all this mean?

What is Intelligence?

The first controversial question that arises from this study is the very definition of intelligence, and specifically how does standard IQ testing relate this intelligence. Cognitive scientists have struggled with this question for decades, and many classic definitions have been criticized for being too narrow, or even biased. In the extreme some argue that intelligence is impossible to measure because any measurement will be overwhelmed with subjective and cultural bias – often based in sexism and racism.

While I do not adhere to this more extreme view, I do think that measuring intelligence is highly problematic. The difficulty stems from the fact that there is no recognized all-inclusive definition of intelligence. This is because intelligence is not one thing, but rather is comprised of many separate neurological functions. It is also because choices must be made in the definition that are not ultimately objective.

For example, we could try to define intelligence as the net effect of many underlying sub-abilities, such as facility with math, logic, memory, perception, emotional maturity, personal insight, ability to abstract, etc. All of these identifiable individual cognitive functions work together to create a combined net effect we experience as “intelligence.”

Or, we could also conceptualize many separate intelligences. We could therefore talk about memory intelligence, math intelligence, and emotional intelligence.

It is now completely clear from neuroscientific evidence that different parts of the brain are dedicated to specific functions, and that individuals can be geniuses in one area but morons in another. The question remains, however, is there a generic intelligence. Do some people’s brains function better, faster, etc. in a way that gives them a cognitive advantage across the board (i.e. not localized to one brain area)?

Nature vs Nurture

In addition to global vs focal intelligence, there is also the question of how much intelligence is hard-wired and genetically determined, and how much is learned. The evidence suggests that there is a strong, perhaps even dominant, genetic component to measures of intelligence. However, evidence and experience also suggests that the more people learn the better able they are to further learn. Also, learning in one area improves the ability to learn in other areas. And finally, remaining intellectually active seems to have a protective effect against developing dementia.

This question has many wrinkles I don’t have time to explore now. For example, is there a genetically determined or acquired desire to learn or basic curiosity that improves intelligence in all areas? Or is ability in each area largely separate?

What is Normal?

There is a movement within the autism community to characterize those who do not have autism as “neurotypical” rather than normal. The implication is that autistics are not abnormal or unhealthy, they are just atypical – different. Personally I think this debate is misguided and largely semantic. It should not distract us from some underlying realities.

The first is what I consider to be the evolutionary perspective – namely that no biological configuration, from an evolutionary perspective, can be considered optimal or normal. All variation is just that, variation. Some people are more outgoing and socially adept, others are more inwardly focused, some are very creative, others are very concrete and detail oriented, etc. Some variations may be more adaptive than others, but that determination can only be made in the context of a specific environment. Change the environment and suddenly a different variation might emerge as more adaptive.

Having said that, not all variation is necessarily equal. Some variants may be superior or inferior in all but the most extreme or quirky contexts. Also, it is reasonable to judge the adaptiveness of cognitive variants in the context of existing civilization. For example, in a literate society the inherent inability to read or write can reasonably be considered to be a deficit.

With regard to autism, the answer of whether or not autism is a normal variant or a disorder is somewhat complex, and depends largely on the severity of the autism type. In mild cases of autism it seems that a decrease in social ability is exchanged for an increase in math and engineering type skills. I think such mild forms of autism, which may be called Aspergers syndrome, can reasonably be characterized as a normal variant, although one that comes with certain challenges in our current society. Everyone has cognitive strengths and weaknesses, and this is just one pattern.

Severe autism, however, goes beyond just a recognizable pattern of cognitive strengths and weaknesses and, in my opinion, can only be characterized as a disorder. Whatever up-side might be perceived in severe autistics is overwhelmed by very significant impairments.

Measuring Intelligence

With all the above in mind, what is the role for measuring intelligence and can it be reliably done? This depends upon ones’ goals, which should be carefully defined. If the goal is to predict how individuals will perform in certain situations, such as in college or graduate school, IQ tests actually do quite well. Standard IQ is certainly a rough test of overall intelligence, and in fact it measures a few markers of intelligence, which are in fact highly dependent on language but also make measures that are independent of language, such as visual perception skills.

It is understand that standard IQ tests are not he beginning and ending of all things cognitive, but rather they are standardized markers for cognitive ability that have been validated in that they have been shown to have predictive power. So if you have a specific question – can this individual function on their own in society – IQ tests are very helpful.

What IQ tests are not is a definitive and objective global measure of every aspect of intelligence. But don’t be confused – they are neither presented or used as such in the mainstream (of course there is always fringe abuse, even of legitimate technology).

The persistent challenge to IQ tests, however, is that they do not apply equally to all identifiable subgroups. In the past they have been criticized as being biased against blacks, non-Europeans, women, etc. I am personally not familiar enough with the data to render an opinion either way. My sense, however, is that testing has evolved to minimize such biases and that they are therefore still useful, but I would not rely upon such testing to measure small differences between groups, because you can’t be sure what you are actually measuring.

Now people with autism are being identified as a subgroup for which the standard IQ test is not an appropriate measure of cognitive ability. The new test looks more at abstraction, which is relatively preserved, and perhaps even above average, in autistic individuals. This is interesting, and teaches us something about the nature of autism. It also reinforces what is not surprising at all – that autistics (like everyone) have more ability in some areas than in others. But I don’t think we can say that one test is objectively “better” than the other. They are just different. Which test is more appropriate depends, again, upon what your goals are.

One side note: in this blog entry I referred to people with autism as “autistics, “autistic individuals,” and “people with autism.” I think that all such terms are perfectly reasonable and respectful. I know that some people take offence at one term or another, but since there is no consensus and I am bound to piss off somebody regardless of which term I use, I might as well piss off everybody. The bottom line is that I don’t think people should go out of their way to take offence at terminology as long as it is reasonably accurate and respectful.