What’s the highest possible IQ?
--
First, an awful clickbait ad:
I’m not clicking through to play your stupid game. The smartest person in the world doesn’t have an IQ of 300.
Just looking at statistics of the bell curve, how high should we expect the numbers to go?
The smartest person in the world is probably Chinese. China is the most populous country in the world, with 1.38 billion people. With that many people, someone should be 6.05 standard deviations above the mean. China also has a higher average IQ than most countries, around 105. Add those together and the smartest person in China should have an IQ of 196.
But the smartest person might also be Jewish. Ashkenazi jews tend to score higher on intelligence tests than other groups, getting an average IQ around 115. There are only about 23 million Jewish people in the world and they’re not all Ashkenazim. Assuming they all have the same mean IQ, the most intelligent person should be 5.35 standard deviations above the mean, for an IQ of 195.
Despite being only 0.3% of the world’s population, Jewish people have won more than 25% of Nobel prizes in physics, medicine, and economics. The Manhattan project drew heavily on Jewish immigrants from Europe. Having a higher average IQ pays off heavily, when it comes to people at the high end of the distribution.
The smartest person in the world is probably a man. That’s not because men are smarter than women — both genders seem to be equally intelligent, on average. But there’s more variability among men, more men are outliers that are either very dumb or very smart. At an IQ of 140, men are overrepresented, 3 to 2. And the same at an IQ of 60.
There’s reason to think this difference is biological. This pattern is consistent in test scores across 40 different countries. The pattern shows up at a young age, perhaps as young as 3 years old. And variability is higher among males, in other traits, and in other species. This seems like a feature of evolution — evolution experiments more with males. Men compete, women pick the winners.
One estimate is that men have 14% higher variability in intelligence than women do. Adding that in, we might guess that the smartest man in China has an IQ somewhere around 202 and the smartest woman around 190.
Again, high variability also hurts some men on the low end of the curve. If there’s someone in China that’s 6.05 standard deviations above the mean, there should also be someone who’s that far below, with an IQ around 8.
I’m not sure what it would mean to be that dumb or how you’d test someone… I don’t think they’d be able to solve even the simplest of puzzles:
I suspect the entire concept of intelligence breaks down, below a certain point. The measured bell curve is not symmetrical, there’s an extra bump of people on the left side:
The bell curve is formed by adding a whole bunch of small positive and negative effects together. Some people have low intelligence because many random genes have added up to a low result. A person can also have low intelligence because they have one serious genetic flaw that harms them. This makes up that extra bump of individuals. You can tell the two groups apart by looking at their siblings — a person who is retarded by down’s syndrome or some other major mutation has siblings with average IQ’s. A person that is low IQ without any major mutation has siblings that are also low IQ.
The concept of intelligence also breaks down a bit at the high end. In general, a person that’s above average at math is also above average at reading and writing. When looking at the most extreme cases, though, this relationship breaks down a bit. The best mathematician in the world is not going to be the best writer in the world. The concept of single most talented person is flawed. To be the best at any one thing takes years of work and specialization. Either way, we’re probably looking at an upper limit on IQ somewhere around 200.
If we look up this question, we can find a lot of clickbait lists of famous smart people, trying to identify the smartest person in the world.
There’s some problems with the stats here. The lists often include children, and IQ is sometimes measured differently in children and adults. People get smarter with age, up until maybe their late teens. For younger children, you can just divide “mental age” by “actual age” to get IQ. This concept doesn’t really work above 15 years old.
Marilyn Vos Savant claims an IQ of 228, because she earned a 23 year old test score at 10. She was incredibly precocious, and scored pretty well (170–180's?) on tests as an adult, but it’s unlikely she’s the smartest person to ever live.
Terence Tao is one possible candidate for the world’s smartest man.
There’s probably some kid out there today who will be grow up to be smarter. As the world population increases, we should see a few more outliers: more exceptional athletes like Usain Bolt or Michael Phelps, and also more exceptionally gifted minds.
But the average population IQ is more important. To get one smartest person with an IQ ten points higher (from 202 to 212), we’d need a Chinese population of 100 billion or an Ashkenazi Jewish population of 1 billion. An IQ of 300 would still be so rare that we shouldn’t see it, even with a much larger population. To get to 300, we’d have to raise the world’s average IQ. Could we do that? And how high could it go?
I wrote two recent posts on disparate topics, but I noticed there are a couple things in common.
In this post, I talked about how much people have modified plants, to make modern crops. The example I gave was corn, which came from a grass called Teosinte:
A more interesting example would be the domestication of wild cabbage, which was bred into 6 different vegetables that we eat today:
Vox gives the example of breeding of chickens for size. In only 50 years of breeding, we’ve selected broiler chickens to be 4 times as large:
It seems incredible, so I looked up the source paper. This might be more about growth rate than it is about adult size. The 3 chickens in question are 56 days old, the modern breed can grow 4 times as large in that time. Still, it’s an efficient way to feed more people. Similar techniques are being used to enhance the milk and meat output of cows.
And, of course, we’ve historically bred wolves into a wide variety of domestic dogs, from chihuahuas to greyhounds.
In another post, I talked about the possibility that we can select for smarter people, in the future. How much smarter are we potentially talking about?
Couples using In Vitro Fertilization currently make about 10 embryos and then implant one for a pregnancy. We can use genetic screening to pick the one that will probably be the smartest. Of 10 embryos, at least one should be one standard deviation above the mean. If our genetic score correlates 70% with actual intelligence then the best pick should be about 10 IQ points higher than the average. We can already predict height this well, and height and intelligence seem to be similarly heritable.
But that’s just the effect on one generation. Maybe we could go a lot higher, over time.
A variety of the enhancements we’ve made on plants and animals increase a trait by 30 standard deviations from the original mean. A 30 SD improvement in intelligence would be an IQ of 450. Just for reference, a 30 SD change in human height would make a person about 16 feet tall.
Steve Hsu looks at the question from another angle. He estimates the number of genes that all have a small positive or negative effect on intelligence and how much each one likely contributes. Then he asks, simply, “what would happen if you turned on every positive gene and turned off every negative one?”
His estimate ends up with an IQ over 1000.
So… who knows if he’s right? Maybe some of those genes lead to redundant effects. There might be side effects. There could be negative things that are correlated with intelligence, like mental illness or autism.
But maybe there will be no big trade-offs at all. Evolution has selected for a lot of things, like disease resistance. It’s not obvious that it has optimized for raw intelligence. High human intelligence already comes with costs. Human infants are helpless for years, compared to other animals. Our large brains use a lot of calories. Infant’s heads are large, compared to a woman’s pelvic bones, making childbirth dangerous for women. Today we have C-sections, we have copious amounts of food, and we’re not competing with other predators on the African savannah. Bigger brains might not be a liability at all.
My post on extinction also included a map of human migration patterns.
There are some interesting details here. People moved into Australia by 50,000 BC, but New Zealand wasn’t populated until much more recently, by Polynesian sailors.
My post on education talked about the controversial theory that IQ varies between racial groups. If you give an IQ test to these two groups, the results are strikingly different. The Maori score fairly high, around 90, while Aborigines score shockingly low, around 65.
Native Hawaiians, also the descendants of Polynesian sailers, score around 85, pretty close to the Maori. All 3 places were later colonized by white people, the native populations suffered in each case, but what really seems to matter isn’t colonial history. What matters is the genetic history of where people came from.
Mainland Africa has been populated with humans for a million years. 250 miles away, the island of Madagascar wasn’t settled until Polynesian sailors moved there. The modern population is some mix of Polynesian and African genes. I can’t find much data on IQ testing in Madagascar, the biggest study I can find tested 150 people who scored around 82, about 10 points higher than mainland Africans did at the time. This would be interesting to study further. In general, it should be easy to compare genetic ancestry with test scores, to better understand how much of human differences come from genetics and how much from discrimination (In both Australia and New Zealand, indigenous people make up big percentages of the prison population).
Why talk about all this?
People have been debating nature vs. nurture for a very long time. Some people on both the left and the right believe in a blank slate. The left thinks that inequality comes from privilege and oppression and we could make poor people succeed if we just had enough government intervention. The right thinks that poor people could succeed if they just tried harder.
In fact, some people are unlikely to ever succeed in life, because they have some disability. Low intelligence is a such a disability. The US military gives cognitive tests to all recruits and sets limits on intelligence. 80% of the military needs to have an IQ above 92. 20% of recruits can have a lower IQ, down to about 81, but only if they graduated from high school. Anyone below that level is considered to be unsuited for any job in the military.
What happens if we assume people are all the same, and recruit less intelligent people? We tried this experiment in Vietnam. Defense secretary Robert McNamara tried relaxing the army’s limits to get more soldiers to fight. These recruits were later called McNamara’s morons. They were 3 times as likely to get killed in action. They were more likely to be reassigned and retrained than other soldiers. And the “morons” who came back alive ultimately did worse than similar (low IQ) men who didn’t serve, going on to earn less money and get divorced more often. Perhaps they were permanently stressed by the experience?
The one group that regularly talks about genetic differences between groups is the alt-right. People sometimes get drawn into this viewpoint when they see information the left never talks about. If we ignore a topic, or tell blatant lies, the alt-right is the side that gets to control the narrative.
Over time, the left is making many ideas into unspeakable taboos. Earlier in this post, I mentioned that men’s IQ scores vary more than women’s. In 1997, Malcolm Gladwell could write about gender and variability in the New Yorker. In 2005, Larry Summers lost his job at Harvard for talking about the same fact. In 2018, activists blocked 2 mathematicians from even publishing a paper about this idea.
Steven Pinker tried talking about this problem. The left suppresses some common truths about the world. He picked 4 examples of topics that the left is hesitant to discuss:
•capitalist societies function better than communist ones
•men and women differ in their life priorities and their tastes and preferences
•some ethnic groups commit violent crimes at much higher rates than others
•islamic groups commit a large majority of worldwide terrorist attacks
Pinker suggested that a young person that sees facts like this for the first time might be persuaded to join the alt-right and accept a more right wing narrative on these issues. The left responded, of course, by calling Pinker a member of the alt-right.
In principle, though, we could also have a hereditarian left, that recognizes differences between people and between groups and still has a desire to help people that need help. In fact, I think we’ll have to get to this point, as automation increases and more people are excluded from the workforce. Right now, the army sets an IQ threshold around 81. Suppose that in the future, automation makes it hard for anyone to get a job if their IQ is below 90. Now we’re talking about 26% unemployment. We could make a plan to support people, as they get laid off, rather than pretending everyone’s the same, and that West Virginia coal miners will all just become coders.
Most importantly, we could also take our modern understanding of genetics and try to improve humanity. In his famous 1969 paper, “How Much Can We Boost IQ and Scholastic Achievement?”, Arthur Jensen argued that we have very little ability to increase IQ through education or preschool. This gave a rather bleak outlook — we have big differences between people, perhaps between groups, we spend billions a year trying to fix these gaps, we stress daily about the differences that remain, but have no remedy.
With a small amount of genetic research, though, we could give the next generation an average IQ of 110, instead of 100. That’s about the difference between making it easy for everyone to finish college vs making it easy to finish high school. Bernie Sanders wants to give everyone free college. I’m okay with that idea, but I would take it one step further — I want everyone to be smart enough to benefit from college.
It should be possible to go a lot further, over time. In a few generations, kids could all be as smart as Einstein. We’re busy arguing over whether one group is maybe 10 or 15 points ahead of another, when we could be working towards a future where people eventually have IQ’s around 200 or 500 or 1000. Think of the discoveries we might make in science and medicine. Think of what that might mean for the future of humanity. We might finally be able to win expert level Sudoku.
