Please place all off-topic comments for the week March 11 to March 17 here.  They will not be posted in regular articles.

[PLEASE PLACE ALL OFF-TOPIC COMMENTS HERE.  THEY WILL NOT BE POSTED IN THIS THREAD]

Some readers have been very critical of the thesis presented by the book Why Only Us by Noam Chomsky and Robert C. Berwick.    Berwick jokes that Chomsky deserves only 5% of the credit for the book.

The thesis seems to be that language developed very recently and rapidly, in either a single, or a small number of genetic mutations.  Commenter Mug of Pee condemned this theory saying:

VIQ is heritable and varies continuously. the very lowest VIQ humans are no smarter than chimps at language.

imagine that the first anatomically modern humans were verbal idiots…. then the utility of language selected for higher and higher verbal ability, language evolved. and the same thing would happen in any large brained species with a vocal apparatus similar to that of humans. homo erection spoke, just really badly.

the idea that a single mutation produced the language faculty is 100% AUTISTIC.

First of all, humans with the linguistic ability of a chimp are probably always those with organic mental retardation, meaning disabilities caused by the overriding effect of single mutant genes or chromosomal abnormalities.  You never see anyone with familial retardation (i.e. low IQ caused by the extreme end of normal variation) who resembles a chimp linguistically.  This is important because it suggests that the difference between man and ape may not be “continuous” as Mug of Pee suggests, but discrete.

In other words, what Chomsky and Berwick are arguing is that prior to about 175 kya or so (give or take perhaps 125,000 years), humans not only had less verbal ability, but a different kind of verbal ability.  In other words they seem to think language is almost a binary ability.  You either have it or you don’t.

Specifically, if we cloned a human from 175,000 years ago, and taught them English, they could acquire a vocabulary almost as big as ours, but if we said to them: “bird ate worm”, they might picture this in their mind:

In other words, they would understand all the words of the sentence, but they would forever be separate concepts.  They would never be able to see the big picture.

By contrast, after the genetic mutation(s) occurred, when you say “bird ate worm” they would picture this in their mind:

The difference between these two interpretations is what Chomsky calls “merge”, the ability to glue different concepts together to form a bigger concept.  And then once you merge multiple concepts into one, you can then merge the multifaceted concepts themselves together to create an even more complex thought “bird ate worm at night” giving rise to almost infinite levels of creativity.

If I understand Chomsky and Berwick’s thesis correctly, this capacity to merge ideas is what makes human verbal ability not just quantitatively different from pre-human communication, but qualitatively different too.

Of course all this is based in part on the somewhat discredited idea of “the great leap forward”, the notion that cultural progress in ancestral humans occurred in a sudden largely upper paleolithic explosion, rather than a gradual transition over millions of years.  More and more scientists believe in gradualism, leaving Chomsky and Berwick in the minority.

But this “merge” ability Chomsky describes extends beyond just language.  According to Berwick, humans are the only animal that makes multifaceted tools, for example a pencil with an eraser at the end.  With one side you write and with the other side, you correct mistakes.  Apparently the two men believe such tools were impossible to conceive before “merge” mutated into existence.

Here’s Chomsky talking about “merge”:

[PLEASE PLACE ALL OFF-TOPIC COMMENTS HERE.  THEY WILL NOT BE POSTED IN THIS THREAD]

In the below video Noam Chomsky talks again about when language likely emerged.  He notes that anatomically modern humans (AMH) are about 200,000 years old and claims there’s no evidence of symbolic behavior predating AMH.

He then claims that the San split-off from other AMH about 150 kya, and since they have language, Chomsky seems to be hinting that language emerged sometime between 200 kya and 150 kya.  Perhaps 175 kya is a good guess.

There are a few potential challenges to Chomsky’s argument:

1. Some scientists are now suggesting AMH might be 325,000 years old, not 200,000 years old.
2. According to Greg Cochran, the split between Bantu and African pygmies was 300 kya, and since both these groups have language, this suggests language must be at least 300,000 years old (unless gene flow since the split explains why both groups have it). Although Peter Frost disagrees with Cochran.
3. An engraving likely made by Neanderthals found in a cave in Gibraltar (and more recently, representational cave drawings of animals) implies AMH may not have been the only species capable of symbolic behavior. If Neanderthals also had this ability it suggests they may have inherited it from the common ancestor they share with us, implying symbolic behavior (and thus language) may be at least 700,000 years old.

[PLEASE PLACE ALL OFF-TOPIC COMMENTS HERE.  THEY WILL NOT BE POSTED IN THIS THREAD]

We’ve talked a lot on this blog about Noam Chomsky, and also about Richard Klein, a scientist who argued that a single brain mutation about 50 kya sparked behavioral modernity (the rise of art, symbolism, and complex tools).

But what we haven’t talked about, is that Noam Chomsky essentially endorses Klein’s theory, only Chomsky is much more specific: he believes the genetic mutation was the capacity for language.

While a lot of comments in the comment section blather on about Chomsky’s political theories, these same comments are mum when it comes to Chomsky’s linguistic theories.  Why?  Because the latter theory is much more abstract and people lack the cognitive ability to discuss it.

Chomsky believes in part that before about 100,000 years ago, humans could only understand linear language, but because of a mutation after 100,000 years ago, we suddenly acquired an ability to arrange words in stacked triangles, and thus think hierarchically, so adverbs could modify verbs that were linearly far away in a sentence.  So as Chomsky’s co-author Robert C. Berwick likes to say:

“Instinctively, birds that swim, can fly”

Instinctively refers to the ability to fly, not the ability to swim, even though linearly, the former verb is closer.  So if Chomsky and Berwick are right, if we cloned someone from 100,000 years ago and raised them in the best modern schools today, they would still not be able to understand this sentence because their brains were only wired to process linear, non-hierarchical language.

Chomsky cites the complete lack of symbolism in the archeological record prior to 100,000 years ago as evidence that there was no language.

In Chomsky’s view, language mutated in the brain of a single African about 100 kya, but because he or she was the first to possess language, she had no one to talk to and just talked to herself.

But perhaps once she had kids who shared her capacity for language, the entire family could speak among themselves and plan strategies that allowed her to exploit the environment in ways their pre-language contemporaries could not, causing the language mutation to spread rapidly.

Many years ago some scientists decided to make a monkey out of Chomsky, literally.  They raised a chimp as though it were human, and named the Chimp “Nim Chimpsky” and tried to teach it human language.

“The experiment was a total failure” gloated Chomsky.  Apparently, the chimp could only string words together linearly, but lacked the hierarchical processing for true language.

A more recent challenge to Chomsky’s language theory is the claim that Neanderthals had language.  Chomsky finds it completely unconvincing and wonders, if Neanderthals had language, why didn’t they use it?  Chomsky’s analogy: It would be like a wild species bird not knowing it could fly until a bunch of biologists tossed them in the air and said “hey you guys can fly!”

For a more in depth discussion about Chomsky’s views, see this lecture by his co-author Robert Berwick:

[NOTE FROM PUMPKIN PERSON: The following is a guest article by Zeitgeisterfahrer and does not necessarily reflect the views of Pumpkin Person.   PLEASE PLACE ALL OFF-TOPIC COMMENTS HERE.  THEY WILL NOT BE POSTED IN THIS THREAD]

I discovered this image that shows the average faces of of many countries.

The average face of each country looks much better than a random person
from that country. If I had to rate the attractivity of 100 people in 100 rounds against the country’s average face, the the latter would likely win in most cases.

Averageness equals attractivity.
Attractive people are average.
Attractive people represent the average of a country or race.
Their behavior and IQ are more average and their standard deviation is smaller.

What is the reason why we prefer beautiful people?
Natural selection!

We prefer beautiful people because we are the descendents of people who prefered attractive (who we now consider attractive) people and mated with them or were beautiful themselves.
They had more offspring and their offspring prefered people with similar faces.
Those who didn’t prefer them had less offspring.

Unattractive people don’t resemble our ancestors that much.
Their package wasn’t very successful and people who considered them okay or attractive were sorted out.
That’s why we have a preference for some faces and and an aversion to other faces.

On the bell curve, average people are the most numerous.
If average people represent the average of a race and averageness equals attractivity, attrative people represent their race in IQ, body, face and behaviour.

Attractive people are the best representation of our ancestors. Their body and mind shows us, how the abilities or our ancestors were and
how and if we diverge from them.

The most attractive people are the mean of the bell curve of our ancestors.

They are less likely to be very dumb or very smart.

They show us how our generation compares to them.

Especially interesting is the comparison of a whole country vs a selection of attractive people.

According to Satoshi Kanazawa, the average IQ of attractive people (american) is 104,23.
The IQ for men is 105 and for women it’s 103,64.
I don’t know if it’s based on a mean of 100 or 98 (american IQ).

According to my theory, americans are 4-6 points less intelligent than their ancestors.

Selection of course has an effect on the perception of attractivity. That means attractive people are a bit less smart than our ancestors,
because the people who are less average and who have more offspring become more average over time.

To get the most representive IQ our our ancestors, you have to first choose a bunch of attractive people.
But we won’t measure their IQ. They will now choose among a new list of attractive people.
The average IQ of the latter group is the best represenation of the intelligence of our ancestors.

If beautiful people in a society have an above average IQ, that means the society is dysgenic and selects for less intelligent people.

If beautiful people in a society have a below average IQ, that means the society is eugenic and selects for more intelligent people.

It’s highly likely that attractive people are smarter in all western countries, which means that we have become more stupid.

Please place off-topic comments for the week here

[NOTE FROM PUMPKIN PERSON: PLEASE PLACE ALL OFF-TOPIC COMMENTS HERE.  THEY WILL NOT BE POSTED IN THIS THREAD]

When I was tested at age 12, I remembered thinking “how are they going to test my intelligence?  Every question you can ask me reflects not just my intelligence, but the knowledge I have acquired.  How do you create a test of innate intelligence separate from acquired knowledge?”

One of the reasons I fell in love with the Wechsler scales is that when I finally sat down to be tested, I was blown away by one of the subtests.  It was the purest, most culture reduced intelligence test I could imagine.  It required nothing but coloured blocks being manipulated to create abstract designs.

It struck me as such a culture free test, that not only could you travel back 2000 years in times and give this test to Jesus, but you could go back 40,000 years and give it to wild Neanderthals.

Another fairly culture reduced test on the Wechsler involved repeating digits from memory, however this would need to be translated into the language of Jesus and Neanderthals to get good results.

Many people don’t think culture reduced tests are possible.  I remember asking my cognitive science professor in university what she thought of the idea, and she said the very idea of testing is a cultural act, so no test could be culture fair.

While academics find the idea of comparing different human races on “culture reduced tests” to be anathema, they have no problem comparing the intelligence of humans and non-humans on such tests.  For example, a respected study in 2007 by the prestigious Max Planck institute compared “chimps, orangutans and 2.5 year-old children” on a battery of tests and “found all to be about equal in the physical cognitive skills of space, quantities and causality. In the social skills of communication, social learning and theory-of-mind skills, the children were correct in about 74 percent of the trials, while the two ape species were correct only about 33 percent of the time.”

How can intelligence tests be culture reduced enough to compare wildly different species, yet too culturally biased to compare different human cultures?

I think part of the answer is that most of the culture reduced tests used for cross cultural comparison have been paper-pencil tests like the Raven Progressive Matrices, and paper-pencil tests are inherently cultural because they require looking at a page in a culture specific way, and school acquired habits like sitting down and concentrating.

To be truly culture reduced, tests need to have some kind of practical relevance that all cultures can relate to, such as getting food for example, as this test of crow intelligence illustrates:

As Arthur Jensen noted on page 248 of Bias in Mental Testing, more support for culture reduced tests comes from this quote from a psychologist who gave a lot of Performance type IQ tests to Kalahari Bushmen.  There was clearly a positive correlation between how well the Bushmen did on these culture reduced tests and how smart the Bushmen considered each other, suggesting the tests were indeed valid for Bushmen.  The Bushmen

accepted as a matter of fact that the “clever ones” would do well on them.  The kind of individual the Bushmen recommended to us, e.g., as a guide when we needed one or as one whose opinion in important matters must be obtained, tended to have above average scores on our tests.  The Bushmen’s concept of “practical intelligence” does not appear to differ essentially from ours (Reuning, 1972, p. 179)

[NOTE FROM PUMPKIN PERSON: PLEASE PLACE ALL OFF-TOPIC COMMENTS HERE.  THEY WILL NOT BE POSTED IN THIS THREAD]

I’ve posted about the IQ of Homo Erectus before, but in this post, I will refine my analysis based on new information.

Brain size of Erectus

According to research cited by scholar Richard Lynn, Homo erectus emerged 1.7 million years ago with an average cranial capacity of 885 cc and by 200,000 years ago, their cranial capacity had increased to 1,186 cc.

How does this compare to modern white cranial capacity?  Anthropologists still cite data from before WWII which is useless because people were short and malnourished back then.  The most up to date direct data I could get on white brain size was a 1980 study of autopsied brains at the Case-Western Reserve University’s Medical School in Cleveland, Ohio.  At age 25 white male brains averaged 1570 g and white females averaged 1339 g,  which I converted to in vivo volumes of 1492 cc and 1273 cc respectively, for a sex-combined white mean 1383 cc.

People use the terms brain size and cranial capacity interchangeably, but the reality is that much of your intracranial volume is cerebrospinal fluid cushioning your brain from blows to the head.  Indeed when I looked at the brain volume/intracranial volume ratio for control subjects in table II of this study,   I found it to be 0.906, which means the sex-combined white brain volume of 1383 cc I cited above, needs to be divided by 0.906 to get the sex-combined cranial capacity, which is 1526 cc.  And the within race, within sex, standard deviation for cranial capacity appears to be about 91 cc ( see table 1 of Rushton’s data), which means that when Erectus first appeared with their 885 cc heads, they were 7.04 SD smaller brained than today’s whites, but by the end of their run when they were 1,186 cc, they were 3.74 SD smaller brained than today’s whites.

The 0.45 correlation between IQ and brain size suggests that for every 1 standard deviation increase in brain size, IQ is expected to increase (on average) by 0.45 standard deviations,  so when Erectus was 7.04 SD smaller brained than today’s whites, their expected IQ was 7.04 SD(0.45) = 3.17 SD lower than modern whites.  Since whites by definition have a mean IQ of 100 with an SD of 15, this gives early Erectus an expected IQ of 52.

However towards the end of its run, Erectus was only 3.74 SD smaller brained than today’s whites, implying an IQ that was only 3.74 SD(0.45) = 1.68 SD lower.  In other words an IQ of 75.

How well does brain size predict the fossil evidence for cognitive ability?

The IQ of Homo erectus 

Experiments suggest that it’s not until a child is seven that she has the mental capacity to create the kind of stone tools Homo erectus created. In other words, Homo erectus may have had the intelligence of a Western seven-year-old. On the WISC-R IQ test, an incipient adult (age 16.9) who performs like a seven-year-old on the spatial construction subtest scores lower than 99.5% of biologically normal members of his generation. In other words, an IQ of about 60.

But we should keep in mind that the research on seven-year-old tool making ability was published in 1979. Probably because of better nutrition/health, truly culture reduced spatial skill has been improved by about 0.2 points a year until 2006 (when U.S. nutrition gains seem to have ended). So Homo erectus probably had an IQ around 55 on the most recent culture reduced Western norms (U.S. white norms).  This is about the IQ predicted from the brain size of early erectus, however it’s well below the IQ 75 predicted for bigger brained later erectus.

Either erectus had other brain deficiencies that prevented their IQs from rising above the 50s despite increasing size, or they made more sophisticated tools later in their existence that I’m not aware of.

[NOTE FROM PUMPKIN PERSON: PLEASE PLACE ALL OFF-TOPIC COMMENTS HERE.  THEY WILL NOT BE POSTED IN THIS THREAD]

Before investigating the IQ of chimps, it’s interesting to ask what the expected IQ chimps should have given their brain size.

Chimp brain size predicts an IQ of 20

The average chimp has a cranial capacity of about 400 cubic centimeters.  The average cranial capacity of white adults in modern Western countries is not known because most of the cited figures come from people who lived before WWII and we know brain size like height has increased since then.  Perhaps the best estimate for modern white cranial capacity is 1474 cubic centimeters, since this is the cranial capacity of Eurasians before the malnutrition and disease of agriculture shrunk our brains and bodies, and with modern nutrition, we’ve hopefully recovered this lost brain size in the generations since WWII!

Given that the within sex standard deviation for cranial capacity among whites is about 91 cc, this implies the chimps are 11.8 standard deviations smaller in brain size than modern whites.  Given about a 0.45 within sex correlation between brain size and IQ in Western samples, this predicts their IQs would be 11.8 SD(0.45) = 5.31 SD below the white mean.  Given that the (white) American mean and SD for IQ are defined as 100 and 15 respectively, this equates to an IQ of 100 – 5.31(15) = 20 (white norms)

How do chimps score on “IQ tests”?

In 2007 there was a fascinating study that compared human 2.5 year-olds to chimps and other apes on a battery of intelligence tests.  With the exception of social intelligence, where the human toddlers were way ahead, the apes and toddlers had the same intelligence.

In other words, chimps have the same intelligence as a 2.5 year old (white) human.

What adult IQ does a mental age of 2.5 equate to?  The question is a lot trickier than it seems.  One could define adult mental age as 16+ and then use the age ratio method to conclude that since 2.5 is 16% of 16, a mental age of 2.5 equates to an adult IQ of 16.  The problem with this method is it assumes intelligence develops as a function of age in a linear way, which is an oversimplification.

What is needed is an actual intelligence test that’s been given to both adults and to toddlers and one where scores increase on an interval scale.

One such test is digit span.  Since the earliest days of intelligence testing (digit span has virtually no Flynn effect) it’s been known that by the age of three, a white child can repeat two digits, which probably means a 2.5 year old can repeat one digit.

By contrast U.S. adults average a forward digit span of 6.645 with a standard deviation of 1.35 and since races in the U.S. differ very little on forward digit span, this should be taken as the white adult distribution.  This means that an adult who performs like a 2.5 year-old (digit span of 1) is  4.18 standard deviations below the white adult average.

If we assume that most cognitive abilities are like digit span,  then chimps (who score like 2.5 years olds on most tested cognitive functions) perhaps average 4.18 standard deviations below white adults on the average test.

Does this mean their IQs average 4.18 SD below the average white adult?  No, because if you score 4.18 SD below white adults on the average test, your composite score on a battery of tests is actually much lower.  Why?  Because it’s much more rare to average an extremely low score across a battery of tests than it is  to score that low on any one test.  Indeed based on the intercorrelation of WAIS-IV subtests, someone who is 4.18 SD below average on the average subtest, would be 5.73 standard deviations (86 IQ points) below average on the composite score, thus my best guess for the average IQ of chimps is 14 (white norms).

This is remarkably close to the predicted IQ of 20 based on chimp brain size.  The fact that the IQ of a species separated from us by over 6 million years of evolution, can have their IQs so closely predicted by the same regression line predicting Western human IQ from Western human brain size, suggests that it’s possible to create culture reduced IQ tests that transcend not only culture, time and space, but species too.

IQ is measuring something very real and biological.

[NOTE FROM PUMPKIN PERSON: PLEASE PLACE ALL OFF-TOPIC COMMENTS HERE.  THEY WILL NOT BE POSTED IN THIS THREAD]

An Israeli reader asked me to convert Israel’s Psychometric Entrance Test to IQ.  According to Wikipedia:

The Psychometric Entrance Test (PET, colloquially known in Hebrew as “the Psychometric“—ha-Psikhometri, הפסיכומטרי) is a standardized test in Israel, generally taken as a higher education entrance exam. The PET covers three areas: mathematics, verbal reasoning and the English language. It is administered by the Israeli National Institute for Testing and Evaluation (NITE) and is heavily weighed for university admissions.

The test may be taken in Hebrew, Arabic, Russian, French, Spanish, or combined Hebrew/English. There are generally five dates the test may be taken each year, between February, April, July, October, and December. Hebrew may be taken at any date; Arabic at four dates; Russian and combined Hebrew/English at two dates; and French and Spanish at one date. Taking the test on two consecutive dates is not allowed; this results in the test being disqualified. The results are valid for university admission for seven years.

Wikipedia provides the following percentiles for the PET

Of course the above percentiles apply only to the subset that takes the PET and not the general Israeli young adult population.  In order to know how select this group is, we need to first know how many Israelis become university age each year (even if they never attend university).

Based on stats from this article, I was able to infer that about 179,000 babies were born in Israel in 2015.  Assuming this is typical of how many are born there each year for the past few decades, then 179,000 become old enough to take the PET each year.  Of these, how many make it through high school and then go on to take the PET?  According to my Israeli reader, only 70,000 take the PET each year which means that the top 2% of test takers (those scoring 725+ overall and 145+ on the subsection) reflect 1400 people.

If we assume that 100% of the most academically skilled Israeli youth take the PET and whatever shortfall there might be is balanced by top foreign talent, then those top 1400 scores reflect not just the highest among the 70,000 test takers, but the highest among all 179,000 Israelis that age.  In other words, the top 0.78% of the general Israeli population.

What IQ does this equate to?  On a scale where the white mean and SD are defined as 100 and 15 respectively, the general U.S. mean and SD are 97 and 15.5.  Israeli’s mean IQ is reportedly 94 (103 for Israeli Ashkenazi, 91 for Israeli Sephardics and 86 for Israeli Arabs) and given it has an ethnically diverse population like the U.S., it’s SD is likely also as high as 15.5.  The normal curve predicts that in a country with a mean IQ of 94 and an SD of 15.5,  the top 0.78% have IQs of 131 (white norms).

Thus:

PET overall score of 725 = IQ 131 (U.S. white norms); 133 (U.S. norms)

PET subsection score 145 = IQ 131 (U.S. white norms); 133 (U.S. norms)

The next question is what is the average IQ of people who write the PET.  About 39% of Israelis of a given birth cohort write the PET each year (70,000 out of 179,000) so PET takers reflect the top 39% of Israeli educational status, which means the median PET taker reflects the top 19.5% of education status.  Normalizing the education distribution, the median PET taker is +0.93 SD in education status.  I don’t know the correlation between IQ and education in Israel, but in the U.S. it used to be 0.7, but has sunk to 0.55 in recent decades.  Let’s split the difference and assume it’s 0.63 in Israel.  This implies PET takers who are +0.93 SD in education would be +0.93 SD(0.63) = 0.59 SD in IQ.  And assuming an Israeli IQ mean and SD of 94 and 15.5 respectively, this equals 0.59(15.5) + 94 = 103.

Thus:

median overall PET score 538 = IQ 103 (U.S. white norms); 106 (U.S. norms)

median subsection PET score 108 = IQ 103 (U.S. white norms); 106 (U.S. norms)

Now that we’ve estimated the IQ equivalents for both the median and the high extreme of the PET distribution, we can tentatively deduce the following linear equations (assuming the relationship is roughly linear):

IQ (white norms) = 0.1497(overall PET score) + 22.44

IQ (U.S. norms) = 0.1444(overall PET score) + 28.32

IQ (white norms) = 0.7568(subsection PET) + 21.27 

IQ (U.S. norms) = 0.7297(subsection PET) + 27.19

Because of ceiling bumping, I suspect these formulas greatly underestimate IQ equivalents as you approach the test ceiling.

According to my Israeli reader “most universities, at least the famous ones in Israel require 740 to enter medicine, 700 computer science.”