Fourmilog: None Dare Call It Reason

« HotBits: Server 3.9 released, JSON support | Main | Your Sky and Solar System Live Updates »

Saturday, February 2, 2019

Reading List: At Our Wits' End

Dutton, Edward and Michael A. Woodley of Menie. At Our Wits' End. Exeter, UK: Imprint Academic, 2018. ISBN 978-1-84540-985-2.

During the Great Depression, the Empire State Building was built, from the beginning of foundation excavation to official opening, in 410 days (less than 14 months). After the destruction of the World Trade Center in New York on September 11, 2001, design and construction of its replacement, the new One World Trade Center was completed on November 3, 2014, 4801 days (160 months) later.

In the 1960s, from U.S. president Kennedy's proposal of a manned lunar mission to the landing of Apollo 11 on the Moon, 2978 days (almost 100 months) elapsed. In January, 2004, U.S. president Bush announced the “Vision for Space Exploration”, aimed at a human return to the lunar surface by 2020. After a comical series of studies, revisions, cancellations, de-scopings, redesigns, schedule slips, and cost overruns, its successor now plans to launch a lunar flyby mission (not even a lunar orbit like Apollo 8) in June 2022, 224 months later. A lunar landing is planned for no sooner than 2028, almost 300 months after the “vision”, and almost nobody believes that date (the landing craft design has not yet begun, and there is no funding for it in the budget).

Wherever you look: junk science, universities corrupted with bogus “studies” departments, politicians peddling discredited nostrums a moment's critical thinking reveals to be folly, an economy built upon an ever-increasing tower of debt that nobody really believes is ever going to be paid off, and the dearth of major, genuine innovations (as opposed to incremental refinement of existing technologies, as has driven the computing, communications, and information technology industries) in every field: science, technology, public policy, and the arts, it often seems like the world is getting dumber. What if it really is?

That is the thesis explored by this insightful book, which is packed with enough “hate facts” to detonate the head of any bien pensant academic or politician. I define a “hate fact” as something which is indisputably true, well-documented by evidence in the literature, which has not been contradicted, but the citation of which is considered “hateful” and can unleash outrage mobs upon anyone so foolish as to utter the fact in public and be a career-limiting move for those employed in Social Justice Warrior-converged organisations. (An example of a hate fact, unrelated to the topic of this book, is the FBI violent crime statistics broken down by the race of the criminal and victim. Nobody disputes the accuracy of this information or the methodology by which it is collected, but woe betide anyone so foolish as to cite the data or draw the obvious conclusions from it.)

In April 2004 I made my own foray into the question of declining intelligence in “Global IQ: 1950–2050” in which I combined estimates of the mean IQ of countries with census data and forecasts of population growth to estimate global mean IQ for a century starting at 1950. Assuming the mean IQ of countries remains constant (which is optimistic, since part of the population growth in high IQ countries with low fertility rates is due to migration from countries with lower IQ), I found that global mean IQ, which was 91.64 for a population of 2.55 billion in 1950, declined to 89.20 for the 6.07 billion alive in 2000, and was expected to fall to 86.32 for the 9.06 billion population forecast for 2050. This is mostly due to the explosive population growth forecast for Sub-Saharan Africa, where many of the populations with low IQ reside.

This is a particularly dismaying prospect, because there is no evidence for sustained consensual self-government in nations with a mean IQ less than 90.

But while I was examining global trends assuming national IQ remains constant, in the present book the authors explore the provocative question of whether the population of today's developed nations is becoming dumber due to the inexorable action of natural selection on whatever genes determine intelligence. The argument is relatively simple, but based upon a number of pillars, each of which is a “hate fact”, although non-controversial among those who study these matters in detail.

1. There is a factor, “general intelligence” or g, which measures the ability to solve a wide variety of mental problems, and this factor, measured by IQ tests, is largely stable across an individual's life.
2. Intelligence, as measured by IQ tests, is, like height, in part heritable. The heritability of IQ is estimated at around 80%, which means that 80% of children's IQ can be estimated from that of their parents, and 20% is due to other factors.
3. IQ correlates positively with factors contributing to success in society. The correlation with performance in education is 0.7, with highest educational level completed 0.5, and with salary 0.3.
4. In Europe, between 1400 and around 1850, the wealthier half of the population had more children who survived to adulthood than the poorer half.
5. Because IQ correlates with social success, that portion of the population which was more intelligent produced more offspring.
6. Just as in selective breeding of animals by selecting those with a desired trait for mating, this resulted in a population whose average IQ increased (slowly) from generation to generation over this half-millennium.

The gradually rising IQ of the population resulted in a growing standard of living as knowledge and inventions accumulated due to the efforts of those with greater intelligence over time. In particular, even a relatively small increase in the mean IQ of a population makes an enormous difference in the tiny fraction of people with “genius level” IQ who are responsible for many of the significant breakthroughs in all forms of human intellectual endeavour. If we consider an IQ of 145 as genius level, in a population of a million with a mean IQ of 100, one in 741 people will have an IQ of 145 or above, so there will be around 1350 people with such an IQ. But if the population's mean IQ is 95, just five points lower, only one in 2331 people will have a genius level IQ, and there will be just 429 potential geniuses in the population of a million. In a population of a million with a mean IQ of 90, there will be just 123 potential geniuses.

(Some technical details are in order. A high IQ [generally 125 or above] appears to be a necessary condition for genius-level achievement, but it is insufficient by itself. Those who produce feats of genius usually combine high intelligence with persistence, ambition, often a single-minded focus on a task, and usually require an environment which allows them to acquire the knowledge and intellectual tools required to apply their talent. But since a high IQ is a requirement, the mean IQ determines what fraction of the population are potential geniuses; other factors such as the society's educational institutions, resources such as libraries, and wealth which allows some people to concentrate on intellectual endeavours instead of manual labour, contribute to how many actual works of genius will be produced. The mean IQ of most Western industrial nations is around 100, and the standard deviation of IQ is normalised to be 15. Using this information you can perform calculations such as those in the previous paragraph using Fourmilab's z Score Calculator, as explained in my Introduction to Probability and Statistics.)

Of the pillars of the argument listed above, items 1 through 3 are noncontroversial except by those who deny the existence of general intelligence entirely or the ability of IQ tests to measure it. The authors present the large body of highly persuasive evidence in favour of those items in a form accessible to the non-specialist. If you reject that evidence, then you needn't consider the rest of the argument.

Item 4, the assertion that wealthier families had more children survive to adulthood, is substantiated by a variety of research, much of it done in England, where recorded wills and church records of baptisms and deaths provide centuries of demographic data. One study, for example, examining wills filed between 1585 and 1638 in Suffolk and Essex found that the richer half of estates (determined by the bequests in the wills) had almost twice as many children named in wills compared to the poorer half. An investigation of records in Norfolk covering the years 1500 to 1630 found an average of four children for middle class families as opposed to two for the lower class. Another, covering Saxony in Germany between 1547 and 1671, found the middle class had an average of 3.4 children who survived to become married, while the working class had just 1.6. This differential fertility seems, in conjunction with item 5, the known correlation between intelligence and social success, to make plausible that a process of selection for intelligence was going on, and probably had been for centuries. (Records are sparse before the 17th century, so detailed research for that period is difficult.)

Another form of selection got underway as the middle ages gave way to the early modern period around the year 1500 in Europe. While in medieval times criminals were rarely executed due to opposition by the Church, by the early modern era almost all felonies received the death penalty. This had the effect of “culling the herd” of its most violent members who, being predominantly young, male, and of low intelligence, would often be removed from the breeding population before fathering any children. To the extent that the propensity to violent crime is heritable (which seems plausible, as almost all human characteristics are heritable to one degree or another), this would have “domesticated” the European human population and contributed to the well-documented dramatic drop in the murder rate in this period. It would have also selected out those of low intelligence, who are prone to violent crime. Further, in England, there was a provision called “Benefit of Clergy” where those who could demonstrate literacy could escape the hangman. This was another selection for intelligence.

If intelligence was gradually increasing in Europe from the middle ages through the time of the Industrial Revolution, can we find evidence of this in history? Obviously, we don't have IQ tests from that period, but there are other suggestive indications. Intelligent people have lower time preference: they are willing to defer immediate gratification for a reward in the future. The rate of interest on borrowed money is a measure of a society's overall time preference. Data covering the period from 1150 through 1950 found that interest rates had declined over the entire time, from over 10% in the year 1200 to around 5% in the 1800s. This is consistent with an increase in intelligence.

Literacy correlates with intelligence, and records from marriage registers and court documents show continually growing literacy from 1580 through 1920. In the latter part of this period, the introduction of government schools contributed to much of the increase, but in early years it may reflect growing intelligence.

A population with growing intelligence should produce more geniuses who make contributions which are recorded in history. In a 2005 study, American physicist Jonathan Huebner compiled a list of 8,583 significant events in the history of science and technology from the Stone Age through 2004. He found that, after adjusting for the total population of the time, the rate of innovation per capita had quadrupled between 1450 and 1870. Independently, Charles Murray's 2003 book Human Accomplishment found that the rate of innovation and the appearance of the figures who created them increased from the Middle Ages through the 1870s.

The authors contend that a growing population with increasing mean intelligence eventually reached a critical mass which led to the industrial revolution, due to a sufficiently large number of genius intellects alive at the same time and an intelligent workforce who could perform the jobs needed to build and operate the new machines. This created unprecedented prosperity and dramatically increased the standard of living throughout the society.

And then an interesting thing happened. It's called the “demographic transition”, and it's been observed in country after country as it develops from a rural, agrarian economy to an urban, industrial society. Pre-industrial societies are characterised by a high birth rate, a high rate of infant and childhood mortality, and a stable or very slowly growing population. Families have many children in the hope of having a few survive to adulthood to care for them in old age and pass on their parents' genes. It is in this phase that the intense selection pressure obtains: the better-off and presumably more intelligent parents will have more children survive to adulthood.

Once industrialisation begins, it is usually accompanied by public health measures, better sanitation, improved access to medical care, and the introduction of innovations such as vaccination, antiseptics, and surgery with anæsthesia. This results in a dramatic fall in the mortality rate for the young, larger families, and an immediate bulge in the population. As social welfare benefits are extended to reach the poor through benefits from employers, charity, or government services, this occurs more broadly across social classes, reducing the disparity in family sizes among the rich and poor.

Eventually, parents begin to see the advantage of smaller families now that they can be confident their offspring have a high probability of surviving to adulthood. This is particularly the case for the better-off, as they realise their progeny will gain an advantage by splitting their inheritance fewer ways and in receiving the better education a family can afford for fewer children. This results in a decline in the birth rate, which eventually reaches the replacement rate (or below), where it comes into line with the death rate.

But what does this do to the selection for intelligence from which humans have been benefitting for centuries? It ends it, and eventually puts it into reverse. In country after country, the better educated and well-off (both correlates of intelligence) have fewer children than the less intelligent. This is easy to understand: in the prime child-bearing years they tend to be occupied with their education and starting a career. They marry later, have children (if at all) at an older age, and due to the female biological clock, have fewer kids even if they desire more. They also use contraception to plan their families and tend to defer having children until the “right time”, which sometimes never comes.

Meanwhile, the less intelligent, who in the modern welfare state are often clients on the public dole, who have less impulse control, high time preference, and when they use contraception often do so improperly resulting in unplanned pregnancies, have more children. They start earlier, don't bother with getting married (as the stigma of single motherhood has largely been eliminated), and rely upon the state to feed, house, educate, and eventually imprison their progeny. This sad reality was hilariously mocked in the introduction to the 2006 film Idiocracy.

While this makes for a funny movie, if the population is really getting dumber, it will have profound implications for the future. There will not just be a falling general level of intelligence but far fewer of the genius-level intellects who drive innovation in science, the arts, and the economy. Further, societies which reach the point where this decline sets in well before others that have industrialised more recently will find themselves at a competitive disadvantage across the board. (U.S. and Europe, I'm talking about China, Korea, and [to a lesser extent] Japan.)

If you've followed the intelligence issue, about now you probably have steam coming out your ears waiting to ask, “But what about the Flynn effect?” IQ tests are usually “normed” to preserve the same mean and standard deviation (100 and 15 in the U.S. and Britain) over the years. James Flynn discovered that, in fact, measured by standardised tests which were not re-normed, measured IQ had rapidly increased in the 20th century in many countries around the world. The increases were sometimes breathtaking: on the standardised Raven's Progressive Matrices test (a nonverbal test considered to have little cultural bias), the scores of British schoolchildren increased by 14 IQ points—almost a full standard deviation—between 1942 and 2008. In the U.S., IQ scores seemed to be rising by around three points per decade, which would imply that people a hundred years ago were two standard deviations more stupid that those today, at the threshold of retardation. The slightest grasp of history (which, sadly many people today lack) will show how absurd such a supposition is.

What's going on, then? The authors join James Flynn in concluding that what we're seeing is an increase in the population's proficiency in taking IQ tests, not an actual increase in general intelligence (g). Over time, children are exposed to more and more standardised tests and tasks which require the skills tested by IQ tests and, if practice doesn't make perfect, it makes better, and with more exposure to media of all kinds, skills of memorisation, manipulation of symbols, and spatial perception will increase. These are correlates of g which IQ tests measure, but what we're seeing may be specific skills which do not correlate with g itself. If this be the case, then eventually we should see the overall decline in general intelligence overtake the Flynn effect and result in a downturn in IQ scores. And this is precisely what appears to be happening.

Norway, Sweden, and Finland have almost universal male military service and give conscripts a standardised IQ test when they report for training. This provides a large database, starting in 1950, of men in these countries, updated yearly. What is seen is an increase in IQ as expected from the Flynn effect from the start of the records in 1950 through 1997, when the scores topped out and began to decline. In Norway, the decline since 1997 was 0.38 points per decade, while in Denmark it was 2.7 points per decade. Similar declines have been seen in Britain, France, the Netherlands, and Australia. (Note that this decline may be due to causes other than decreasing intelligence of the original population. Immigration from lower-IQ countries will also contribute to decreases in the mean score of the cohorts tested. But the consequences for countries with falling IQ may be the same regardless of the cause.)

There are other correlates of general intelligence which have little of the cultural bias of which some accuse IQ tests. They are largely based upon the assumption that g is something akin to the CPU clock speed of a computer: the ability of the brain to perform basic tasks. These include simple reaction time (how quickly can you push a button, for example, when a light comes on), the ability to discriminate among similar colours, the use of uncommon words, and the ability to repeat a sequence of digits in reverse order. All of these measures (albeit often from very sparse data sets) are consistent with increasing general intelligence in Europe up to some time in the 19th century and a decline ever since.

If this is true, what does it mean for our civilisation? The authors contend that there is an inevitable cycle in the rise and fall of civilisations which has been seen many times in history. A society starts out with a low standard of living, high birth and death rates, and strong selection for intelligence. This increases the mean general intelligence of the population and, much faster, the fraction of genius level intellects. These contribute to a growth in the standard of living in the society, better conditions for the poor, and eventually a degree of prosperity which reduces the infant and childhood death rate. Eventually, the birth rate falls, starting with the more intelligent and better off portion of the population. The birth rate falls to or below replacement, with a higher fraction of births now from less intelligent parents. Mean IQ and the fraction of geniuses falls, the society falls into stagnation and decline, and usually ends up being conquered or supplanted by a younger civilisation still on the rising part of the intelligence curve. They argue that this pattern can be seen in the histories of Rome, Islamic civilisation, and classical China.

And for the West—are we doomed to idiocracy? Well, there may be some possible escapes or technological fixes. We may discover the collection of genes responsible for the hereditary transmission of intelligence and develop interventions to select for them in the population. (Think this crosses the “ick factor”? What parent would look askance at a pill which gave their child an IQ boost of 15 points? What government wouldn't make these pills available to all their citizens purely on the basis of international competitiveness?) We may send some tiny fraction of our population to Mars, space habitats, or other challenging environments where they will be re-subjected to intense selection for intelligence and breed a successor society (doubtless very different from our own) which will start again at the beginning of the eternal cycle. We may have a religious revival (they happen when you least expect them), which puts an end to the cult of pessimism, decline, and death and restores belief in large families and, with it, the selection for intelligence. (Some may look at Joseph Smith as a prototype of this, but so far the impact of his religion has been on the margins outside areas where believers congregate.) Perhaps some of our increasingly sparse population of geniuses will figure out artificial general intelligence and our mind children will slip the surly bonds of biology and its tedious eternal return to stupidity. We might embrace the decline but vow to preserve everything we've learned as a bequest to our successors: stored in multiple locations in ways the next Enlightenment centuries hence can build upon, just as scholars in the Renaissance rediscovered the works of the ancient Greeks and Romans.

Or, maybe we won't. In which case, “Winter has come and it's only going to get colder. Wrap up warm.”

Here is a James Delingpole interview of the authors and discussion of the book.

Posted at February 2, 2019 16:08