As announced late August, here is the promised discussion with Davide Piffer on ancient genomics: where we are, and where we hope to go. 1 hour and 7 minutes.

Transcript

Transcript generated by AI, might have errors.

0:00:14 – Emil

Hello YouTube, we’re trying something else this time. Here we are in beautiful Como, in Italy, and instead of me blabbering alone, we will be two people blabbering today for your enjoyment. Over here, we have a Davide with us and is native Italy, in fact. Well, not too close to home, but close enough from my perspective. Recently, we published an ancient Rome study where we looked at the recent released ancient Roman genomes which were published by some academics and which you can download freely, and, together with Davide, who is sitting over here, we’ve analyzed these to test the ancient Rome eugenics eugenics, collapse of civilization hypothesis. And maybe, davide, you want to summarize the study, since you were the main author.

0:01:10 – Davide

Yeah, this study relied on genomic samples from different periods from central Italy. So they start from the Copper Age, bronze Age and then they go all the way from the Iron Age, which is also like partially overlaps with the Republican period, followed by the Imperial period, late antiquity, medieval period, and then I also managed to find a modern sample. And this genomic sample allowed us to test this secret theory of civilizations where the level of intelligence rises and causes an increase in social complexity and civilization, but when the civilization becomes too comfortable, this will relax the selection pressures and cause some disgenic effects, causing a lowering of the polygenic scores. And so we were able to test this model and this we actually found like in line with our the predictions from our theory, that there was a huge spike in the polygenic force for educational attainment and cognitive ability from the Neolithic Copper Age and Bronze Age to the Iron Age. This increase was about two standard deviations and then it was followed by a decrease. This decrease was not the same magnitude, it was a bit smaller in magnitude to the increase. So the level, the cognitive level, didn’t go back to the baseline if we take the Neolithic periods as the baseline, but it was about two thirds of the way and after the Imperial period there was another increase in late antiquity and the Middle Ages. This actually could also be interpreted as a confirmation of our theory, because the late antiquity and the Middle Ages were times of unrest and war and when the civilization collapsed and the cyclical model of civilization predicted hard times will actually have a new genetic effect.

So this is what we actually found. But, to be honest, this later increase was not statistically significant. So the statistically significant effect was for the lower polygenic score in the pre-Iron Age sample and the higher one for the Iron Age Republican sample. So we have to admit that the Iron Age sample was small, was only 11 people and this out of the total sample was 127. So the results needs to be taken with caution. The more reliable effect is the increase from the pre-Iron Age Neolithic sample to later ages, which is a line what we find with genomes from other parts of Europe and even East Asia. There was in prehistory and throughout history there was a eugenic trend. So there was an increase in the polygenic, the genetic scores for education and cognitive ability. So when we will get more data we will be able to test whether the linear model of an increase of the polygenic scores is a better fit to the data compared to a cyclical model.

0:05:39 – Emil

So that’s the main summary of our study. Looking at our questions we got from the readers. Of course the criticism is mainly about the small sample of the Republican period or the Iron Age, which is only 11 people. But it’s not really that something we could do about, because, as far as we know, this was the only Roman ancient genomics available for study, and so this should be seen really as a tentative first try at this, and when we find more data at some point it will get sequenced. Ancient genomics is booming. Thousands of samples, I think, come out every year.

0:06:16 – Davide

We already have some related samples from the Etruscan civilization which are very similar, very related to ancient Romans, because the Etruscans were absorbed into ancient Rome during the Iron Age and these were published recently and we can actually test our model using these as proxies for our Roman sample.

0:06:44 – Emil

Yes, and that’s for this study.

The next viewer question is from Peter Frost and he’s interested in Greg Clarke’s hypothesis, which is that in his book the farewell to arms he hypothesizes that, based on records of marriage and death certificates wills to when someone dies.

You can see that during the Middle Ages in England and maybe Scotland and these in the United Kingdom you can see that the number of surviving offspring or the surviving number of children that the wills were left to had a positive correlation with the amount of money in the wealth of this family. So in other words, it used to be the case throughout medieval times, up until around 1900, 1850, depending a bit on exact location and method, that there was a positive selection for wealth at least and for education and occupation status. And so, since these are free, well known correlates and effects of intelligence, we would expect that in this period Britain, or the UK at least, was under eugenics selection in the same way that we see with the ancient Roman data and, as we’ll talk about later, with the other ancient, ancient Chinese data and this and Peter Frost, the main question is whether we have enough ancient data from Western European sources to test this Clarkeian theory in general. Do we have other data, medieval UK?

0:08:23 – Davide

data we do. I’m currently imputing medieval British data from the early to intermediate middle ages and it’s a substantial sample of over 450 individuals, so it will enable us to test Gregor Clarke’s theory. The only problem would be that if Woodley’s theory of this genesis writes and there was a huge decrease in England in the intelligence in the last century we would. If we use the contemporary genomes we won’t be able to see. We really have to somehow take into account the potential decrease of intelligence. So it will be. We should also try to get some old samples from the beginning of the 20th century that I guess the UK Biobank data has got older people born in the 30s. So this will be a better sample and this will actually enable us to test the theory Gregor Clarke’s theory that there was something.

This effect was stronger in England because from the data it doesn’t look like it’s the case, because the modern polygenics for the UK are not higher than other European countries and according to Gregor Clarke’s model they should be higher because this trend was stronger in England. But Gregor Clarke’s model could still be correct if we found that a strong disgenic trend in the last century from the Victorian times to contemporary people. There is some I already have some suggested evidence of the decrease from the Middle Ages in northwest Europe, because I’ve got increased, sorry, from the Middle Ages, because I’ve got almost 400 Viking genomes and these are about once half a standard deviation lower than contemporary Western Europeans. So there was, if we assume, of course it’s a strong assumption, if you assume that the Vikings said the same, the Vikings were from the Middle Ages, at the same level of intelligence as the British, then yes, we have, or even other Europeans, then we have some evidence supporting Gregor Clarke’s theory.

0:11:17 – Emil

Kind of a recent selection. So how would the if the Vikings were not mentally superior, at least according to genetically and speaking, of course, as a former Viking, or for my ancestors anyway? So how would the Vikings able to conquer this?

0:11:35 – Davide

do this seem to have any other traits, as I recall the or taller, or something in the yeah the polygenic course for a Viking is significantly higher than the modern, even modern, northern Europeans, at least modern British and Finnish. So it’s possible that physical strength was a help them conquer other and populations and also, like some, navigation skills and other cultural traits or even psychological personality traits that we don’t know.

0:12:16 – Emil

Because I recall the ancient history or not ancient, but the historical sources generally described Vikings as quite tall, and this is what you found also with the Viking.

0:12:26 – Davide

Yeah, and this was a good test of the accuracy of the polygenic of the imputation and the polygenic scores, because the polygenic scores actually found confirm these historical sources which I think he generally has been previously ascribed to.

0:12:44 – Emil

What is it? The medieval warming thing, which I think was unusually strong, and Scandinavia. But maybe, maybe there was a genetic aspect to this Viking superiority in martial arts or in combat, at least for some time. The other question that that Peter has, and some multiple other readers have, is that whether there’s any G-wases for these other components of Clark’s hypothesis. So he’s not just saying that there was an increase in intelligence, but also, say, a slower life history or time preference decreased, or maybe violence, genomics.

0:13:25 – Davide

Well, education G-wases is not purely cognitive, it also includes some life history traits like conscientiousness and slow life history. So it’s a it’s kind of an omnivorous measure, for it’s even probably it’s even a better measure to test Clark’s theory. But we also have, like, a other measure of schizophrenia which showed an opposite trend to the education polygenic score, and there is it’s possible that because we know that schizophrenics tend to be much more violent than the general population, so it’s possible that selection against violence caused the decrease in the polygenic scores for schizophrenia.

0:14:15 – Emil

But currently there, as I recall, there are no, there are not like G-wases of violent crime itself, so you couldn’t look directly for this one. I don’t think there’s any G-wases for time preference either. These traits, though they have a large importance to many people in social science broadly, that doesn’t seem to be enough interest in the genomics community to make sure that they include measures of these.

0:14:40 – Davide

There is a candidate gene studies of the serotonin J gene, the. You know, they repeat yeah, but. Cig yeah, but it’s only one gene and it’s not very useful for comparing populations, because we know that these traits are polygenic. It could give us like an idea, but yeah we know it’s differs. African-americans have a lot higher risk based on this gene.

0:15:12 – Emil

I think there was a math analysis of these CIG repeats and whether it’s actually predicted testosterone or some of these other things, which also I think is related to the 2G digital ratio, not the digit span yeah, and this did not really show much of either of these.

So so maybe this is one of these early candidates genetics things that won’t work. Or maybe once the whole genome sequencing comes out from the UK Biobank, we’ll be able to to calculate the whole genome estimates of the CIG and see if this correlates. I don’t think UK Biobank has any violent crime measure, but they must have some other social status that must be a proxy for it to some extent. Yeah, so this would ever be an interesting one to do.

The next question, also from multiple people, is concerning what you were relating to before with the height, and some people ask whether we are at the, if you can estimate the, the genotypic or say the phenotypic level of intelligence just from the genetics and and the problem with this is that intelligence in a person is a phenotype.

It’s the realized aspect that involves both the environmental input and the genetic input. So if you only have, if you only look at them, one of these over time, you cannot be sure that this is the one that’s driving the trend in the phenotype. So, for instance, the in the ancient Romans they had quite, quite high and genotypic intelligence at least in this small initial study we did, but it’s not really possible to say how this might relate to modern levels of IQ. If you were to naively put current west europeans at 100, supposing we had some perfect italian tests, we couldn’t really say, based on the potentially higher polygenic scores of the ancient Romans, that these would actually be smarter than us, even if they have higher polygenic scores, because it’s possible that our environment improvements are so much better to outweigh this.

I think you would agree with this, maybe I agree with this so unfortunately it’s not possible to tell only from genetics when the peak of human intelligence was, or are we currently living in the peak human intelligence? It’s all these flun debates and the measurement problems also for ancient greek.

0:17:41 – Davide

We got the, for instance, golden estimates, but these were based, I think, on I think on the eight Greeks from Athens, not the general Greek population. The preliminary data we have from ancient greek doesn’t seem to back up his hypothesis, but it’s based because we have data from all over Greece and also greek, so it’s not comparable to the population of Athens. It’s possible that at the time there was some selective migration into Athens or some local eugenic trends what about comparing the ancient Greeks that you had so far?

0:18:22 – Emil

how do they compare to modern Greeks? Is there data there? Yeah, it’s.

0:18:25 – Davide

Uh, there was actually. I compared them to modern Greeks and the modern Greeks are a bit lower than western europeans, modern western europeans and they. There was a very small increase from ancient Greeks to modern Greeks, but that’s because the modern Greeks are not so.

0:18:45 – Emil

So in Greece there was not a strong eugenic trend yeah, it’s a bit difficult to say from these, from these kind of data. Yeah, the next interesting one, as you already remarked upon, is that there is a question about the prevalence of mental illness among different ethnic groups and even the same population over time, and so there are some g wass and schizophrenia, bipolar alchemy or Alzheimer’s and so on, and so his question is how do the ancient samples compare to the moderns, with the apologetics for these mental problems?

0:19:17 – Davide

yeah, the ADHD is not the HD was, as conflicting results and because there is some signal but it’s very small and doesn’t reach the significant levels. I think it’s because it’s not necessary, because there was no selection for the genes related to ADHD, but it’s probably because the g wass is not very powerful and indeed the the predictive power of the ADHD g wass polygenic score is very low. But the schizophrenia polygenic score is a better predictive ability and we find we find a strong effect effect size for it. Basically, all the ancient populations that they tested.

0:20:09 – Emil

They have much higher polygenic risk or schizophrenia compared to the their modern counterparts so one could even speculate, if you were to go quite far here, that the, that ancient humans were more deeply religious, since there’s a known association with schizophrenia and religiousness which, if you put on your atheist glasses and think of religion from a natural perspective, you would say that religious is about over-perceiving, over-perceiving agency in inanimate objects or ascribing agency to, to kind of random things like the river flooding or earthquakes or volcanoes and this sort of thing. And we suggest the ancients were genetically more inclined to ascribe, like the, the coming, the ebbing and the flow of, say, the Nile to some river, god or yeah, not just religion, not just not just religion but also magic.

0:21:11 – Davide

Like the ancient that widely believed in magic and the magical thinking is one aspect in modern, in the modern, our modern rational culture is an aspect of the schizophrenia of the diagnosis. So it’s, it’s likely that they were somehow beneficial in the more primitive societies. The schizophrenia alleles because if they help people make sense of their world, either through magic or through religion, and indeed we we find in modern populations, even among modern populations, that we find like a big gap between different ethnic groups in the schizophrenia risk score so do you recall offhand, if these the polygenic scores of the current world populations that there is genetic data for?

0:22:10 – Emil

do these relate to the official prevailances like the government’s report, or like non-profits estimate?

0:22:19 – Davide

not very well because, for example, the african polygenic score is for schizophrenia is very high, several standard deviations higher than the east asian or european one. But the prevalence, the official prevalence of schizophrenia in africa is not that high. But that’s probably because of the different diagnostic criteria. But if we compare different racial groups living in the same country for example, I’ve looked up studies of black comparing blacks and whites in sweden, in the netherlands, in the us blacks have higher rates of schizophrenia and also much higher polygenic scores for schizophrenia so it seemed that once you kind of equalize the environment and the measurement, then the, the polygenics, are more in line, the polygenic scores are more in line with the, the phenotypes we observe.

0:23:20 – Emil

How about the east asians? How do they rate here?

0:23:22 – Davide

the east asians have very similar levels to europeans, which is also it’s not. It’s in line with the epidemiological data we have from western countries where they they are pretty similar in the schizophrenia rates as to whites okay if we were to be extra speculative.

0:23:48 – Emil

We you know in in in the psychiatry you can talk about the p-factor, which is kind of this general index of mental, mental problems, the general psychopsychopathology factor, and but you can also break this down to certain other dimensions if you say there’s an overall trench where most any diagnosis gives you an increased chance of any other diagnosis maybe not exactly all of them, but very close to all of them.

But you could also say that in some ways schizophrenia is the opposite of autism, where autism is characterized by an under-perceiving agency that exists.

Where artists are very bad at picking up human cues, people looking at them in strange ways and they don’t notice this as as actually reflecting agency. Where schizophrenia would be over-perceiving agency, like kind of taking slides where someone looked at you in the funny way maybe that’s because they hate you. Now you need to be angry at them, and autism is like you would under-perceive these things so you wouldn’t spot the danger coming if someone is sleeping with your wife and these two people are looking at each other a bit too happily. So it would seem that that maybe autism is has a negative effect in this sense and schizophrenia could have a positive sense. It helps you perceive and there would be a an optimal level on this autism schizophrenia dimension. So, based on the ancient data that you told us about, it would seem that everybody is heading towards more autism. This could be considered consistent with our increased emphasis on science and technology and not using anecdotes but doing randomized, controlled trials and kind of formal deductive reasoning, as opposed to this kind of intuition approach.

0:25:41 – Davide

Yes, yeah, the decrease in the schizophrenia genetic predisposition could be due to a selection for, or it’s actually a mirror, and we don’t know which way the causal effect goes. It could mirror our tendencies for rational thinking and to be less superstitious than our ancestors.

0:26:10 – Emil

The next question concerns the cold winter’s theory. The cold winter’s theory for those not familiar with it watching this the cold winter’s theory is an attempt to explain the worldwide variation we see in average intelligence levels. Where you see, generally, if you take country populations as they are in current countries and you take these as a proxy for people who evolved in these locations, which means you have to exclude the new world, which for most is settled and now are inhabited by Europeans mixed with Native Americans and some African slaves later. But if we focus on the old world, so everywhere from West Africa and West Europe all the way over to Asia, but not including the Americas, these are in all cases, a few of them are settler countries, those of them are non-settlers. So you would have about 120 countries with roughly original stock of population. And if you take the various climatic measures of these countries, you can take average summer temperatures, average summer colds like the lowest altitudes, all kinds of these climatic factors, and think of it as a kind of biologist who would approach it.

So if you’re a biologist, you come out and see all these kind of subversions of species or breeds of some new fish and you record all the climates in the different places these fish live. You try to figure out what seems to be driving the evolution here, and so when you do this with humans, what you get from comparing the climate data is that one factor stands out above all the others and it’s the temperature of winters and the low temperature of winters, the cold winters. So the origin of this theory it’s not really that someone sat around with armchair and came up with this came out of. Let’s take a biological perspective and look at all the climatic factors and this one is empirically the strongest one I know.

I’ve looked into some modern data to see was this maybe a coincidence of some of these older? I think the first study on this was published in 2000s, sometime by Alicaba and Templar. But I’ve tried the modern IQ estimates for countries and modern temperature data and this didn’t change the results. It was still by far the strongest one. So the question of the viewers is that there are some obvious anomalies with this model. So, for instance, now we are in Italy, how does the cold winters theory explain the fact that the first advanced white civilizations, or just civilizations in general, all arose in the temperate Mediterranean climate? So maybe you have a guess.

0:28:58 – Davide

Yes, my guess is that the cold winters were at a positive, a new genetic effect.

There were a positive selective pressure on intelligence when the status of civilization was not advanced.

So during, for example, the Stone Age, when the biggest threat to life and reproduction was natural phenomena or predatory animals or disease, and in this, or even keeping yourself warm In this situation, in this environment, the cold climate, would require higher intelligence to survive. But once people developed civilizations and they managed to shield themselves from the natural environment not completely but to a large extent the selection pressures from the social environment became stronger. So the social selection, gene-cultural evolution, became stronger and social complexity became the main driver of evolution. So that’s why there was a positive feedback loop whereby more complex societies selected for more intelligence and then more intelligent selected for even more complex societies, up until the point where, probably in the ancient Rome, there was the opposite trend. But another objection to this theory is that what I recall is that what is more important is the seasonality, not actually the raw temperature. So it’s also probably extremely. If you go too far north, like, for example, the Inuit, their climate is not enough variable, so they probably didn’t have enough pressure on their skills.

0:31:32 – Emil

There are many potential ways, but what I think is the most important fact to recognize with the cold-windows theory is that it’s not supposed to be a theory that explains everything. It’s not a theory of every human group’s absolute high ranking or rank order in the intelligence. It’s supposed to be the main theory. So let’s say the cold-windows could maybe explain 70% of the group differences we see over time. Sometimes there are anomalies.

0:32:04 – Davide

Native Americans are an anomaly because they evolved in Siberia. They split from the East Asians in Siberia so they were in a very cold climate, Also the Inuit, but these people don’t have an extremely high level of intelligence.

0:32:24 – Emil

No, but they’re also not particularly low either. So from this perspective, you could say the Inuit is an outlier, but you could say Europeans are an outlier. Maybe we’re too high for our temperature.

0:32:36 – Davide

Think about the East Asians. There is not a significant difference between the South Chinese and the North Chinese. The Mongolians live in one of the coldest places in the world, but their IQ is not as high as that of the Chinese who live in warmer climates from Hong Kong or southern China. In this case, you can explain this because China was much more densely populated for a much longer time, so the social competition created selective pressure. Overall, there is evidence for the importance of cold winter theories. For example, I found a moderate positive correlation between latitude which is not far from the equator you are, which is a good proxy for cold climate, and the polygenic scores for education, although this relationship is not linear because it seems to peak at some intermediate latitudes. For example, the Yakuts don’t have, despite living in a very cold climate, they don’t have very high polygenic scores.

0:34:07 – Emil

The Yakut IQ, as I recall from some recent Russian studies, is about 95 or so slightly below the Russians.

Considering that they live in colder climates, what we assume ancient Russians or East Slavs used to do. They would seem that they should be smarter, but they can be the counteracting factor that caused that. The population density was very small. It is difficult to spread. If you get a new mutation or you have a mutation, this mutation has a spread for the population. When your infected population size is quite small, then this slows down evolution.

You could say the same thing for the Inuits, where farming was never done by the Inuits. In Greenlands. The Vikings tried farming in Greenland and they died out after some time when the climate changed a bit. But the Indians, the Inuits, never really farmed. They’ve been living essentially like hunter-gatherers. It’s possible that the transition between these different forms of subsistence for humans changes the selection speed or even sometimes the direction.

My favorite pet hypothesis that has so far not been tested, but I will set out the anecdote or not the anecdote, but the evidence I could think of so far which is that I’ve noticed that the Balkans and people who live in South Slavs are quite much lower than the other Europeans who live at the same latitude and approximately the same heat.

So Italians are much smarter than Albanians, even at the same latitude, and some similar effects. One reason for this could be the altitude that there’s a mountains here, and mountains would also make your population have a lower population density and theoretically slow down evolution for this or maybe change the selection pressure. And there are some other cases of this in the Caucasus area, which is also very cold and very bad winters. The people who live in the Caucasus area are much lower intelligence and do an expect based on the climatic, the climate data, even though these are Caucasians I mean, they give origin to the word Caucasian, so they are also an anomaly. And the final example I found is Tibetans in China, I think, have IQs in the 85 or their bouts, roughly the same as African Americans.

0:36:45 – Davide

Peruvians.

0:36:48 – Emil

There’s some data on the Latin America where the people who live higher in the mountains have higher IQs than the people who live closer to the shore. This evidence is ambiguous. It can be self-selection. Maybe the people in the mountains are just more Amerindian, or there’s been selective migration south, which seems likely. Nevertheless, there is some circumstantial some of this evidence that points in this mountain theory of intelligence, if we should call it that. So this is something I hope to look into in the future, if we can somehow figure out a data set where we have different ethnic groups and where approximate altitude they evolved at, no matter where they currently live, just approximate altitude. I guess Nepalese would be an interesting example. The next question is about let’s see this one we talked about already with the Galton’s claim about the ancient Greeks and about in the many thousands of genomes that are coming out. When we can expect to have a kind of final verdict on this Galton Greek hypothesis?

0:38:04 – Davide

The pace at which these genomes are being ancient, genomes are being released, publishes very high. So expecting the next couple of years we will be able to even so like two free years or so.

0:38:16 – Emil

Yeah, I would agree with this. His next question is if we think, or if we agree with Galton’s estimate, that the ancient Greeks would be, comparably to roughly 130 or 115 or so.

0:38:31 – Davide

At least the ancient Athenians. The ancient Athenians or from this particular the golden age of.

0:38:38 – Emil

I mean it is possible. It brings back to this typical discussion of genius, where what Galton’s method consisted of? Taking the population density or the population estimate that he could surmise based on the data he had available in Victorian times, and then the number of Greek geniuses in this time according to historical records, and you can divide these to get the genius rate. And if you compare this to, say, the modern, the Victorian age that he was living in, which is also one of the highest, or the Italian, the North Italian Renaissance or the Dutch Renaissance, and these rates, the Greeks had the highest rates of geniuses. And if you’re willing to make certain normal distribution assumptions with this, you can of course compute the mean IQs, which is what Galton did. The objection to this kind of estimate would be that it was much easier to be a genius in ancient Greek.

Nothing is discovered. You’re just kind of starting up with a blank slate. No hanging fruit, low hanging fruit right. Aristotle, you know, sits around, thinks very hard about logic, comes up with some patterns of logic that makes sense. He has a book on the animals right which consists of kind of a broad A lot of knowledge was borrowed from other countries.

0:39:59 – Davide

There was no copyright, no very Our authorship. A lot of knowledge was transmitted through schools, not in a written form, and a lot of texts were anonymous, so it was very easy to just copy some text from Alexander’s library or some Babylonian text and just claim it’s your own text. Plagiarism was not an issue like it is now, so it’s difficult to compare the rate of creativity at the time to our modern rate.

0:40:43 – Emil

So the final here would be the. If only we had the library of Alexandra, we could check whether the ancient Greeks were really supreme geniuses or supreme plagiarists. It would be a funny one. I have another question about the cold winters here, but this is the one we already covered about the mountain and the Albanians and the Bulgarians, some of these. The next question here is a kind of interesting one related to your current AI worries, or AI hopes, depending on which side of the debate you’re on. We have a rat king asking at what point did humans peed cognitively? Are we at our peak now, or is there some speculation that the Victorians were smarter than us? Or the ancient Greeks maybe? When is a peak human cognitive performance, you think?

0:41:35 – Davide

Well, there is not a, Apparently it’s like. It’s difficult to say because it depends on the place. So different places have different peaks. So, for example, in the case of Italy it could be the Roman Republic or the Renaissance. In the case of England it could be the Victorian age. So it depends on the area. But in general the peak tend to be modern time, because there was this trend for higher intelligence continued throughout modern times. So probably my guess would be it was like in the 19th, early 20th century in general, and then it was followed by this genetic trend.

0:42:29 – Emil

So the question here would be the more complicated one of how to answer the Flynn effect worries with measurement, because of course we now know that from biobank genetic data that the polygenic scores have certainly declined in modern times. We know this from the UK Biobank, we know this from the Icelandic study and I’m sure there will be some other studies soon showing us.

0:42:51 – Davide

Then we have like the reaction time data showing us.

0:42:55 – Emil

Some reaction time data. But to be fair, of course, we have reaction times that get slower in multiple countries. It used to be the Woodley-Metz analysis, based on the Victorian comparison, but there is also a longitudinal Sweden study showing that even Swedes in the 70s 1970s I think have slower reaction times or faster reaction times than the Swedes in the 1990s. So it would seem that this reaction time decline is definitely for real, which would be kind of the opposite of what you’d think, based on the increased prevalence of modern technology, or computer games often has a reaction time component. It would seem like, from a training perspective, that humans should be increasingly trained to have higher reaction time, and yet they keep going down.

0:43:40 – Davide

Yeah, also the driving. The driving emphasis is on reaction time. So if everyone almost everyone has to learn how to drive nowadays, and to be a good driver, you need to. You need to have fast reaction times.

0:43:57 – Emil

But on the other hand we could say that there are a lot of standard IQ tests that we have been measuring intelligence with since the 1905 or so when the Benet first tests came around. And if you look over time with these more complicated tests or complex mental tests like vocabulary size or ability to tell people kind of abstract similarity between two objects, the similarities tests, or these like matrix reasoning tests like the Ravens, these tests show some of them show not just a small increase but like a 30-point increase.

0:44:36 – Davide

The possibilities that the reaction time reflect more like a hardware component, whereas these tests reflect kind of like a software, like an improvement in our software due to our environment. So our software got better, but there was like a due to the environments we live in, but there was a decrease in the genetic predisposition.

0:45:06 – Emil

It’s a one speculative way of doing this?

0:45:09 – Davide

Yeah, that’s my idea.

0:45:11 – Emil

From my perspective, I think the fluid effect is mysterious and no one really knows what’s going on. Some tests go up, some tests go down. You can rate things by abstract level, as people used to think, or still think to some extent, that nonverbal tests are particularly pure measures of general intelligence, of GG factor. But then the primary, like in the nonverbal test, the Raven test, shows one of the largest increases, which doesn’t make any sense from this I’m sure we said the old school kind of a humanitarian perspective. One of the other ones that show basically no trends is the ability to do basic arithmetic, which would be very mysterious. If people have gotten smarter over time, as you would think from the Raven test scores or similarities tests, then why can they not do basic?

0:46:05 – Davide

arithmetic. They didn’t have calculators, so they had to do.

0:46:09 – Emil

No, these are all results from non-calculator data right.

0:46:13 – Davide

No, but because now people in school, when they get homework or tests. They can just use a calculator but in the old times they had to do it by heart or pen or pencil or paper.

0:46:28 – Emil

You could make a similar ad hoc theory for plausible hypothesis, as we call it here.

0:46:36 – Davide

This is my ad hoc explanation.

0:46:39 – Emil

For digit span which, as I recall, also shows roughly no trends. Maybe forward digit span is getting a bit better in some meta-analysis and the backwards digit span is getting a bit worse. So you could come up with some complicated speculation where, well, since people use more calculators now, they don’t need to memorize a lot of numbers in the head like they used to do when they had to do long division in your head or on paper, whereas people could go like what’s 12 plus, you know, plus 12 times 7, like then, like calculated right before they would turn out the.

0:47:13 – Davide

Exactly so. This reliance on calculators can have an impact on our ability to do mental arithmetic. We could test this by also observing the measuring the trend in other tasks, mental tasks that have been impacted by technology. We could draw a list of them and then compare their trend across time, based on how much they were impacted by technology, and see if there was an interaction.

0:47:49 – Emil

It’s just that I think there’s a quite limited set of cognitive tests that we have long data on, like obviously we have the old Stanford Binnett test, but these were given to kids.

So, for instance, if you see that people are getting higher scores over time on the Stanford Binnett, maybe this isn’t adult intelligence increasing, maybe it’s just children mature earlier and so they’re okay. So we’ve kind of sped up developmental speed, but not necessarily increased intelligence of adults. But, as I recall, there’s multiple meta-analysis of the Flynn effect which mostly show that the increase seems to be roughly constant everywhere, even across ethnic groups. In the US the speed is the same, which is difficult to explain with some of these environmental theories where, let’s say, access to computers helped you train your cognitive skills playing games or just like using programming or whatever it is people doing computers and this would train your various skill sets, your reasoning abilities, resulting in these increased results in the IQ tests. But when you see the rate of improvement is roughly constant across wealthy and less wealthy ethnic groups within the US, it makes it difficult to come up with a consistent explanation for this fact.

0:49:11 – Davide

You mean, like, because these different groups have a different rate of technology? Adoption so this seems to be no.

0:49:22 – Emil

Richard Lin, who unfortunately died recently. He pointed out these inconsistencies, if you can come up with all these hypothetical technologies that humans adopted and this helped boost which is kind of typical James Flynn thing, scientific goggles, this sort of spectacles, this sort of thing, as Lin is pointing out, it seems to increase, seems to affect just about everybody.

0:49:45 – Davide

My theory is that we don’t know the origin of the Flynn effect. I think that the guy who actually would be able to explain it should get a Nobel Prize.

0:49:57 – Emil

That is maybe possible. Yeah, but moving from a Nobel Prize to other Nobel Prize, we have one commenter who is interested in learning about the Askenasi history of intelligence. So the most viewers are probably familiar with the Askenasi intelligence of the Jewish, but specifically the German Jews score somewhere between 110-115, depending on which sample went from US Jews, whatever Israeli Jews are a bit lower. So there is a theory by Greg Cochran and what’s the other guy’s called the natural history of Jewish intelligence or Askenasi intelligence, which is that the kind of Jews arose in the Middle Ages or maybe the late antiquity and they were essentially trades people or trade men from Judea, from Israel, and they migrated to Italy, maybe South France, but somewhere in this area, and they took local wives while they were merchants. And since they are merchants, which is a difficult occupation when everybody else is a peasant, the merchant profession would tend to select more for intelligence than being a peasant would, and so on this theory, the Jews would evolve higher intelligence, faster, at a faster rate than the other surrounding populations.

0:51:27 – Davide

This comes back to our theory that social cultural selection can be very strong and trump the natural selection from the climate or the environment in general. And yes, there is some preliminary evidence for this, because we have a small medieval Jewish sample published by the David Reich team, and their polygynous scores for education are lower than modern Europeans.

0:52:03 – Emil

Even so, it seems like the speed of the increase of the polygynous scores among Jews was even higher than among Gentiles which would be the central claim of this model that ideally we would find some ancient pre-European mixed Jews from Israel, but comparable ones, and we would find some of the earliest Jews you can find in Italy or in South France or wherever, and we would keep note their intelligence or the polygynous scores. Over time, and as they kind of move around, they should be coming, initially with some increase, because these are merchants, not peasants, so they should be elite for their origin population. Once they get in Europe they should be self-selecting for eliteness faster or for more intelligence faster than the other Europeans or the Gentiles. And so since we only currently have to my knowledge we only have the Eifer to the medieval German Jews, the problem with this sample is that it’s from the same cemetery, same burial place, so it’s not a representative sample of medieval Jews.

So we need more samples of ancient Jews so that’s our message to the Jewish community watching this, please convince your fellow Jews that we need to sequence more ancient Jews so we can prove your natural superiority.

Jesus, for sure, was not stupid, so maybe there were already some IAQ genes among the ancient Jews moving out from purely our genetic discussion so far, with all the evolution and polygynous scores and ancient samples and sequencing techniques, if we get so far. There is one reader who has a question about the impact of the smart fraction compared to the average intelligence. I don’t know if you have a theory. Of course I would myself. With NorCal, we published a paper on this one year ago, so where we looked into this, and so the main answer to this would be that you can try to estimate the distribution of each population using PISA data. Basically, it’s the only kind of data that has this.

Richard Lin’s compilations of natural IQs have not, in general, been recording the different centiles, and the underlying data is not shared, so it’s not possible to go back and recompute it. You can use the standard deviations themselves, but these are often incomparable or just kind of odd distributions or what is it called restriction of range, issues like ceiling effects, which make the standard errors incomprehensible. But using the PISA data, which is only from 15 year olds or so, you have many years of data. You can calculate how smart is the top 5 smartest students on average from each of these countries. Do that for every year across every test. You can put these on a common scale. You can do the same thing for the bottom 5% student and also for the mean student from a given country. So from this way you can kind of estimate a potentially asymmetric normal distribution where there might be a long tail. If there’s a larger than expected amount of smart people, there’s a long left tail, there’s a larger amount of dull people and, of course, on the smart fraction, theory is that the size of the right tail is important for the development of societies, because these are the people who come up with the new inventions and who are good with studying new businesses and so on.

The problem with this data is that when you take the scores from the PISA, what you find is that the average ability of a country correlates I think it’s 9.7 with the right tail ability, which means that essentially you cannot run a regression on this with the same size of only 100. There’s not enough statistical precision for this. You will get some nonsense estimates because of the extreme collinearity. So we came up with this alternative approach, which is not without its faults. It at least gets you somewhere, which is that for each country you say try to predict the score of the top 5% smarter students and you predict this from the average student and to the degree that the top students are smarter than you would expect based on the average students.

That means that the tail is longer, getting a number that tells you the links of the right tail and you can do the same thing for the left tail. So you get these two parameters telling you about how the normal distribution sort of looks like. And if you do this, then these measures become uncorrelated because you’ve now removed the mean from the other ones. Now they’re uncorrelated, but you can put them in a regression and you won’t get any problems. And if the smart fraction theory is correct, then generally you should see that the smart fraction predicts better outcomes and the dull fraction does not do so to the same extent, or maybe not at all. But what we found in our study was basically what the smart fraction theory says the countries with the larger smart fractions, even holding constant the average intelligence countries, do better on not every outcome you could measure, but most of them would be a p-value of 1% or less and on average it was quite a bit. You could improve it.

0:58:09 – Davide

I have a question because you studied this more in depth than me. What do you think could be the causes of higher variance, of different variances in IQ between countries? Why do you think some countries have higher variance, apart from ethnic composition, like if we are talking about ethnic and homogeneous countries? Why do you think some countries will have higher variance variance in cognitive ability?

0:58:45 – Emil

I would generally think it’s mainly about the ethnic composition, or it doesn’t even have to be ethnic, it could just be take, for instance, like southern and northern Italians. In some sense these are the same ethnic group. In some sense they’re not kind of a continuum.

0:59:02 – Davide

So it’s what geneticists call population structure like a higher level of population structure.

0:59:08 – Emil

I would guess it’s mostly population structure, if not entirely. But if you are to speculate a bit in behavioural, in behavioural genetics terms, say that if you have a high environmental impact in some society then theoretically the, the variance of the IQ should also be higher because you have this additional environmental causation. And if the genetic variance doesn’t somehow decrease in response to this, if there’s not some kind of some kind of strange variance interaction going on, then generally speaking, modern, modern variance in IQ or variability in intelligence should actually been decreasing over time because we’ve we’ve been increasing the environment to be more equalized. I don’t know if it’s possible to compare this over time. This is kind of a secondary Flynn effect, not just as the is the, the mean going up, but the variance should also become, become smaller, and I know the Danish army test shows this pattern. So there’s kind of a the rise of the little guy, as one scholar who studied this set, and I don’t know if this has been generally confirmed, but it would be in line with this.

So but based on this theory you would get that if we look between countries, ethnic groups in them, then essentially poorer countries should have higher variance and therefore to some extent they would also have, to the extent that this variance is not even they might have a strong smart fraction or they might not depends on ethnic subgroups in Africa, of course, many countries have kind of random borders because they were drawn by Europeans who said, well, my colony goes to the end of the this parallel and then down, or so, coming to Africa, africa is an interesting case because we know that Africa is more genetic variation within it than other continents, because that’s this is where, like humans originate, the modern humans originated, and I wonder if this large amount of genetic variance between African populations is also translate translates into higher genetic variance for intelligence?

1:01:42 – Davide

I guess we don’t. I don’t know if we have enough IQ data or polygenic score data to actually test these hypothesis, and we have some populations from a thousand genomes from eastern and western Africa, but from what I could see, there was no big difference between eastern and western Africans from a from the perspective of statistical genetics, if they have more variance in frequencies.

1:02:12 – Emil

But these go in random directions and maybe this wouldn’t affect the standard deviation of a some score like a polygenic score. I’m not sure I had to go test my intuition. There’s a, but it’s a common argument you hear that’s higher. Even some economists are famously based there, like big growth theory, on this kind of reasoning. But I have a blog post arguing that it’s false and I reviewed the various American data from different ethnic groups showing that their standard deviations are mostly the same and they’re in unrelated to the standard deviations. And in fact, if you, if there is any difference, then maybe the African-American standard deviation is a bit smaller instead of larger, because not only is it roughly 80% African which has a higher genetic diversity, but it is also 20% European, which would give an even larger genetic diversity. And yet the standard deviation of this group is either no larger or maybe even slightly smaller, at least if you look at the reason, the ABCD data, which I think showed very close to the same standard deviations.

There’s only one final question, the question before the final question, the final reader question. So it’s about Thomas Saul argued that West Indian blacks greater success in the US disproved explanations that racism or slavery were the main reason for the disparities observed in the US, and so this is his theory that instead they would have a kind of cultural origin and that the, the, the West Indies or the Caribbean blacks, when they came to us, they must have been doing better at the time that Thomas Saul was writing, I think in the 70s or the 80s. This was back when the kind of the great societies was kind of getting on the way. It wasn’t really showing the big improvements promised for, so there were some conservatives like Thomas Saul would come up with like maybe, maybe the theory they, the, the racism of the slavery theory is kind of incorrect, and so it’s really a question for me, maybe because I recently had a blog post.

I’ve replied to the economist, the libertarian economist, david Friedman, and he repeated this Saul argument. So I decided to look up the the modern status of Caribbean Americans, and it turns out the modern Caribbean Americans actually are below average in income. So they are, maybe they’re a bit better than African Americans, but they’re not like much better. They’re maybe slightly better, and so it seems there’s no, there’s no big mystery to explain with the the population of Haiti is not particularly wealthy or educated.

It’s actually one of the most dangerous countries in the world and right the lowest education you would wonder from this perspective, if the Caribbean supposedly brought some great culture with them compared to the African Americans, then why isn’t this culture doing very well back in their home countries? So why do they need to go to America with this? Isn’t it easier to just think that initially, the the people who came from the caribbeans to the US were selective?

but they were positively selected for being more elites, in the same way that we see elite Nigerians and some other, even South Africans of course South Africans are mostly white returnies or sometimes Jews that that come back. So it would seem a much easier to explanation to just explain this in terms of self-selection of immigration, rather than some magical culture effects that doesn’t work at home. So that that was the last the reader question.

1:06:01 – Davide

So if you have any last things you would like to add, I think we covered everything we were asked yeah, maybe tell the reader what the the future holds for your research the next coming year yeah, keep checking my Twitter account for like updates on my publications because there will be lots of new interesting papers coming about ancient genomes and the test of the cyclical model of civilization and Gregorikar’s theory, and the final goal is actually tracking the human population through all over the world through time and space. So if anyone interested in the history, genetics and psychology and human biological differences will find a lot of very interesting material in our publications all right when that sets.

1:07:07 – Emil

I would like to say Skoll to the, to the viewers, and I hope you enjoyed it. Praised frozen.

Transcribed by https://podium.page

0:01:10 – Speaker 2

0:05:39 – Speaker 1

0:06:16 – Speaker 2

0:06:44 – Speaker 1

Yes, and that’s for this study.

0:08:23 – Speaker 2

0:11:17 – Speaker 1

0:11:35 – Speaker 2

0:12:16 – Speaker 1

Because I recall the ancient history or not ancient, but the historical sources generally described Vikings as quite tall, and this is what you found also with the Viking.

0:12:26 – Speaker 2

0:12:44 – Speaker 1

0:13:25 – Speaker 2

0:14:15 – Speaker 1

0:14:40 – Speaker 2

0:15:12 – Speaker 1

0:17:41 – Speaker 2

0:18:22 – Speaker 1

how do they compare to modern Greeks? Is there data there? Yeah, it’s.

0:18:25 – Speaker 2

0:18:45 – Speaker 1

0:19:17 – Speaker 2

0:20:09 – Speaker 1

0:21:11 – Speaker 2

0:22:10 – Speaker 1

do these relate to the official prevailances like the government’s report, or like non-profits estimate?

0:22:19 – Speaker 2

0:23:20 – Speaker 1

How about the east asians? How do they rate here?

0:23:22 – Speaker 2

0:23:48 – Speaker 1

But you could also say that in some ways schizophrenia is the opposite of autism, where autism is characterized by an under-perceiving agency that exists.

0:25:41 – Speaker 2

0:26:10 – Speaker 1

0:28:58 – Speaker 2

Yes, my guess is that the cold winters were at a positive, a new genetic effect.

There were a positive selective pressure on intelligence when the status of civilization was not advanced.

0:31:32 – Speaker 1

0:32:04 – Speaker 2

0:32:24 – Speaker 1

0:32:36 – Speaker 2

0:34:07 – Speaker 1

The Yakut IQ, as I recall from some recent Russian studies, is about 95 or so slightly below the Russians.

0:36:45 – Speaker 2

Peruvians.

0:36:48 – Speaker 1

0:38:04 – Speaker 2

The pace at which these genomes are being ancient, genomes are being released, publishes very high. So expecting the next couple of years we will be able to even so like two free years or so.

0:38:16 – Speaker 1

Yeah, I would agree with this. His next question is if we think, or if we agree with Galton’s estimate, that the ancient Greeks would be, comparably to roughly 130 or 115 or so.

0:38:31 – Speaker 2

At least the ancient Athenians. The ancient Athenians or from this particular the golden age of.

0:38:38 – Speaker 1

0:39:59 – Speaker 2

0:40:43 – Speaker 1

0:41:35 – Speaker 2

0:42:29 – Speaker 1

0:42:51 – Speaker 2

Then we have like the reaction time data showing us.

0:42:55 – Speaker 1

0:43:40 – Speaker 2

0:43:57 – Speaker 1

0:44:36 – Speaker 2

0:45:06 – Speaker 1

It’s a one speculative way of doing this?

0:45:09 – Speaker 2

Yeah, that’s my idea.

0:45:11 – Speaker 1

0:46:05 – Speaker 2

arithmetic. They didn’t have calculators, so they had to do.

0:46:09 – Speaker 1

No, these are all results from non-calculator data right.

0:46:13 – Speaker 2

No, but because now people in school, when they get homework or tests. They can just use a calculator but in the old times they had to do it by heart or pen or pencil or paper.

0:46:28 – Speaker 1

You could make a similar ad hoc theory for plausible hypothesis, as we call it here.

0:46:36 – Speaker 2

This is my ad hoc explanation.

0:46:39 – Speaker 1

0:47:13 – Speaker 2

0:47:49 – Speaker 1

It’s just that I think there’s a quite limited set of cognitive tests that we have long data on, like obviously we have the old Stanford Binnett test, but these were given to kids.

0:49:11 – Speaker 2

You mean, like, because these different groups have a different rate of technology? Adoption so this seems to be no.

0:49:22 – Speaker 1

0:49:45 – Speaker 2

My theory is that we don’t know the origin of the Flynn effect. I think that the guy who actually would be able to explain it should get a Nobel Prize.

0:49:57 – Speaker 1

0:51:27 – Speaker 2

0:52:03 – Speaker 1

0:58:09 – Speaker 2

0:58:45 – Speaker 1

0:59:02 – Speaker 2

So it’s what geneticists call population structure like a higher level of population structure.

0:59:08 – Speaker 1

1:01:42 – Speaker 2

1:02:12 – Speaker 1

1:06:01 – Speaker 2

1:07:07 – Speaker 1

I would like to say Skoll to the, to the viewers, and I hope you enjoyed it. Praised frozen.

Transcribed by https://podium.page

Transcript

You Might Also Like

Measurement error and group differences explained

Intelligence, income inequality and prison rates: It’s complicated

The national IQ of Nepal