It's rare to read a popular science book that's both factually accurate and engaging, but David Reich offers an excellent example of how to do that.

The book explores the newly emerging field of ancient DNA. Using techniques that are less than a decade old, we can now sequence the entire genome of humans up to hundreds of thousands of years old. The picture that's beginning to emerge as we analyse this DNA is shaking up classic views on history and prehistory, and offers some fascinating insights into our story as a species.


The book begins by introducing the techniques and history of the field. At the same time it applies these techniques to studying human population movements and mixtures. It starts off deep in the past and with humanity as a whole, then moves forwards in time to specific areas and peoples. Finally it ends up discussing a little bit about the politics of DNA, the future of the field, and a small amount of philosophy.

Techniques of the trade

DNA might be relatively stable over a human lifetime, but over thousands of years it degrades, such that by the time you're dealing with a neanderthal bone the vast majority of the DNA you'll extract from it is from microbes and other contaminants. The ability to separate any actual neanderthal DNA from it relies on a number of advances:

  1. Improved techniques: for example there's a bone in the ear that preserves DNA at least 10 times better than the rest of the body.
  2. Careful quality control: extracting ancient DNA is impossible if it becomes contaminated with modern human DNA. The bones are taken to a completely sterile room modeled after microchip fabrication plants. There most of the bone is milled away, leaving the center which is likely to be free of human contaminants. It is this remainder that is sequenced.
  3. Powerful computers and advanced algorithms make DNA sequencing thousands of times cheaper than it was just 15 years ago, and make it possible to differentiate human DNA from microbial DNA.

The very first Neanderthal genome was sequenced over 3 years from 2007, and doing this at any kind of scale only became possible from about 2014. At the time of writing his book only a few hundred complete ancient genomes had been published, but thousands more had been recently analyzed and were awaiting publication from his laboratory alone. By now, 3 years after writing his book, over 3000 have been published.

So this is a field where the most exciting discoveries are yet to come. We've begun to explore some of this data, but at the moment the people in this field are very much pioneers, still discovering the shape of the landscape, but not yet ready to start mapping out all the details.

Of course extracting the DNA is only half the problem. You've then got to analyze it. At the moment our understanding of genes is not yet sophisticated enough to learn much about our forbears by directly inspecting their genome, so the analysis in this book is limited pretty much to studying ancestry.

There's a bewildering array of statistical techniques used to do this, and Reich isn't afraid to get into them. I personally greatly appreciated this - it takes a certain amount of skill to tell me the important things I need to know about DNA and statistics to get the ideas across, without telling me so many details that it becomes confusing. However it's easy to skip these passages if that's not your cup of tea.

For example, every stem cell has 2 copies of each chromosome - one from the father and one from the mother. When it splits to produce 2 sperm or egg cells, some of those chromosomes will be spliced at a random location in a process called homologous recombination. So for example, one sperm might have the first 2/3rds of the mother's chromosome 17, and the last 1/3rd of the father's chromosome 17, and the other sperm will have the opposite.

This means that over time what were continuous sections of DNA from a single ancestor will split it up into hundreds or even thousands of tiny fragments spread out across a chromosome.

This can be used to detect how long ago an admixture of two populations occurred. For example European DNA contains about 2% neanderthal DNA. We know the rate at which chromosomes are spliced, so by looking at the average length of the fragments of neanderthal DNA in Europeans we can see that their ancestors bred with neanderthals about 50,000 years ago.

Of Neanderthals and Denisovans

Reich now starts telling us some results that have emerged from the study of ancient DNA. He begins by investigating the origins of Homo Sapiens, going back tens, or even hundreds, of thousands of years.

What I found most fascinating about this chapter is that wherever Homo Sapiens spread they interbred with other Homo species, whether Neanderthals in Europe or Denisovans in Oceania. This helped Homo Sapiens adapt to new climates by mixing with Homo species that were already well suited to the area. Today Neanderthal DNA is very well preserved in European genes for keratin production - keratin is a vital component of hair, so it's possible that Neanderthal hair was better suited for keeping warm in cold climates.

Reich describes how most of the team was shocked to discover that modern Europeans partially descend from Neanderthals. They tried redoing the tests every which way to check if it was caused by some kind of systematic problem in their test procedure, but the results came out the same every time, and was corroborated through multiple lines of evidence. This pattern is repeated throughout the book - some analysis of DNA produces highly surprising results which diverge significantly from consensus, but time after time the results are replicated. Reich is also careful to distinguish fact from hypothesis. Between the two this gives me a very high confidence in most of the results he states.

I was definitely excited to find out I was probably part Neanderthal - this definitely made me rethink how we define humanity, but I can't exactly put my finger on what it changes.

The flow of history

Reich continues by investigating the population flows across history, and across the world, from the many migrations throughout Europe, to the original settling of the Americas and the Pacific islands, to the recombination of two completely seperate populations that make up India as it is today.

All of these are mildly interesting in their own way, but unless you have a love of this kind of thing it can start to get rather dry by the time you're halfway in. Ancient history is fascinating, but mostly when it tells a story. Reich tries to tell a story, but it's somewhat tricky to do that when talking about population flows, with only minimal archaeological evidence to allow us to actually connect to the culture and lives of those he's describing. However he does throw in a lot of anecdotes which tends to make for easier reading, and are often fascinating in their own right.

I found him most engaging when he touched on areas which I knew about from some other source - for example Indo-European languages might have been spread by a group of nomads from the Asian Steppe (the Yamnaya) who invented a culture based on the wheel and the horse which was successful enough to displace much of the existing peoples of both Europe and India.

My guess is that every individual reader will have sections he will find quite dry here, and sections that touch on his interests which he will find fascinating. There's very little dependence between the different sections, so feel free to dip in and out.

For that reason I'm going to skip reviewing this section in detail - instead I'll just highlight the main points covered. Each of these points has an entire chapter or so dedicated to it, so if any interest you I'd recommend buying the book:

  • Evolution of the human species both in and out of Africa for the past hundreds of thousands of years. He challenges the out of Africa theory, suggesting that Homo Sapiens might have left Africa hundreds of thousands of years ago, evolved mostly in  Eurasia, and then migrated back into Africa more recently.
  • Humans interbred with other Homo species on a large number of occasions.
  • Repeated migrations regularly displaced the people in Europe. Most Europeans today are descended from the nomads of the Yamnaya culture.
  • Indians are descended from a Eurasian and an East Asian population that mixed a few thousand years ago. As you go south through India the proportions change gradually from mostly Eurasian ancestry to mostly East Asian ancestry. The caste system in India has led to a large number of extreme population bottlenecks, which leads to high rates of genetic diseases that genetic screening programs (such as Dor Yesharim) can help mitigate.
  • America was settled through a wave of Migrations from Siberia towards the end of the ice age, spreading steadily south throughout the Americas. The different migrations took different routes through the glaciers that covered much of north America, which can still be seen in the ancestry of Eskimo Aleut speakers today.
  • An analysis of what we know about East Asian population movements so far. We have limited data at the moment, which leaves a lot more to be discovered. We are beginning to build up a picture of how the pacific islands and Oceania were settled. This is particularly interesting as the story of human migration and adaptation to islands hundreds of thousands of kilometres apart is incredible, and this begins to give us more insight into how it was done - in theory ancient DNA might be able to build up a picture of the order each island was hopped to, and from where.
  • A study of African history. The out of Africa view has inadvertently distracted attention from modern African prehistory. There's almost a view that humans 'left' Africa, so there's nothing interesting left there to study. He shows numerous recent large migrations across Africa, and shows how that relates to the African language groups. This should put Africa back on the road-map as an area with plenty of rich results to give anyone who studies it. There's also plenty to find further back - for example there's some evidence that two Homo species split deep in the past and then after hundreds of thousands of years of parallel evolution remixed in almost 50/50 proportions to produce modern humans.

The future of Ancient DNA

Reich's last chapter deals with a number of disparate matters, not least of which is the future of the field.

He expects the impact of Ancient DNA on archaeology to eventually be as impactful as carbon dating. Like carbon dating it provides a technique to take what was originally an art (dating archaeological artifacts/plotting ancient population movements) and turns it into a science which is able to give hard answers. Like carbon dating he expects full genome synthesis to eventually become commercialized, with a number of laboratories around the world offering the service for a low fee, and archaeologists routinely sending any human remains they find to these laboratories.

This is certainly a bold claim, and I was skeptical when I read it. However as I'll discuss in the section on the Ancient Canaanites below, it's important not to underestimate the power this field can have to resolve seemingly unanswerable questions.

He also expects the field to become far more specialized. As a pioneer in the field he's got to be a jack of all trades, but master of none - dipping his toes into humans living several hundred thousand years ago one day, and the movements of Polynesians across pacific Islands a couple of thousand years ago the next. He is forced to combine information from multiple disparate fields for his research - genetics of course, but also archaeology, history, linguistics, statistics etc. But no one can be an expert across all these fields and across so many diverse historical populations. In the future as the field expands he expects that researchers will specialize in one particular topic in history, and apply the knowledge that Ancient DNA throws up specifically to their field of interest. This will lead to much better research, but he feels lucky that he gets to have all the fun in the young free cowboy days of the field when he can study whatever he wants.

As a result of these two innovations, he expects full genome sequencing to be applied to a far greater range of problems than it is today. For example he highlights a study looking at how people moved around the British isles over the past 1000 years which provides an extremely concrete setting for anyone who studies British history. We expect a lot more of that kind of really detailed stuff in the future, as opposed to the kind of sweeping studies he's done looking at very high level patterns across vast geographic and temporal distances.

The politics of DNA

Come for the science, stay for the politics.

The pages of this book are littered with political points, anecdotes, and ethical conundrums, which definitely serve to make the book more interesting. I don't necessarily agree with everything he says, but they almost always gave me a lot to think about. Here's a small sample:

Countering Racism

Reich often tries to point out how his research counters traditionally racist views.

For example, he points out how the fault lines between different peoples is completely different to what it was 10,000 years ago - all the distinct 'races' that existed then have since recombined many times over, producing new and different clusters. Hence Race is not a fundamental biological unit, which he feels undermines racist views.

Not being a white supremacist I'm not an expert on racist philosophy, but I'm guessing this is a bit of a straw argument. I could imagine a KKK member simply saying that sure, races change over time, but right now Whites are much better than any other race. I think that giving weak arguments against a fundamentalist ideology is much worse than not giving any at all - "Hah", the KKK member says. "If that's the best argument they've got against me I'm clearly definitely right". I worry that in this instance Reich is more interested in waving his not racist credentials given the controversial nature of some of the results he's published, than actually countering racism.

On the other hand Reich's analysis of race and racism at the end of the book (discussed below) is one of the best I've seen in mainstream literature.

Sensitive Topics

On the other hand there's the story of how his research into Indian ancestry almost couldn't be published, because it suggested that about half of Indian ancestry was West Eurasian - an idea that would be politically explosive in India. In the end they had to re-frame it as Indians are descended from two unrelated groups - North Ancestral Indians and South Ancestral Indians, and South Ancestral Indians are closely related to West Eurasians.

Similarly his research showing that the Corded Ware culture spread via migration rather than cultural spread partially affirmed views that had gone out of favor after the Nazis supported them. This caused a number of researchers to be uncomfortable publishing it - even though it contradicted Nazis views on other aspects.

It's interesting to think how politics and ideology affect the framing of many studies. There's always pressure to understate any conclusions that challenge the reining ideological winds, and emphasize those that agree, and it's important to keep that in mind when reading studies. For that reason it's vital to actually read papers, and not just their abstracts, so that you can see what the study actually says, rather than how the authors wished to present it.

Similarly views can go in and out of favor irrespective of any evidence for them, but because they become associated with a particular ideology. If a the Nazi party were to say that spaces are better than tabs, I'd expect you'd suddenly find all programmers unanimously agreeing that tabs are just obviously better.

Access to Samples

You can't sequence the genomes of bodies you don't have, and  getting access to human remains can often be tricky.

For example museums and research institutions like to keep hold of any human remains they get their hands on. Native american tribes tend to claim the bodies as ancestors of theirs and want the museums to give them back so they can be respectfully buried.

By law museums in the USA only have to do that if the Native Americans can prove that the remains are related to their particular tribe. In practice, if they can show that they're at all related to Native Americans in general (which they invariably are, being before any Europeans came over) the museums don't want to be seen to be making a fuss, and have to strike a difficult balance between holding onto the remains, and letting them go.

This makes getting hold of DNA to study the settling of the Americas quite tricky, and often legally fraught. This also doesn't lead to particular goodwill when asking current Native Americans to donate their blood to genetic research.

Reich highlights the work of Eske Willerslev as someone who is making promising headway. He has decided to work with the Native American community rather than against them - by asking them for permission to extract DNA samples from the remains in order to prove that they are genetically related, and then encouraging the museums to return the remains even if they don't need to under the letter of the law. This gives samples that are useful for research, and leads to better relations with the Native American community.

Genetic Racism

Are some races inherently smarter than others?

Reich has an interesting approach to this question:

  1. There are those that claim that we have evidence that some particular races are genetically more intelligent than others. E.g. Nicholas Wade on Ashkenazi Jews.
  2. Reich believes that any such hypothesis is pure armchair speculation - we have nowhere near enough evidence to be able to say that, and so anyone who is saying that is likely driven by racist motives.
  3. On the other hand there are those who, for equally ideological reasons, claim that inherent differences in intelligence between races is impossible.
  4. They suggest such reasons as that intelligence is driven by a very large number of genes, so it would be unlikely that one race would consistently have more intelligent genes than another.
  5. Reich believes this is nonsense - there are many large and obvious racial differences that are driven by a large number of genes - e.g. height.
  6. We don't yet know if/which races are more intelligent than others, but it is very likely that we will find such differences in the future.
  7. Instead of burying our heads in the sand we should accept that someone's intelligence does not affect their inherent humanity or value, so that we can safely assimilate such information in the future.
  8. In fact we should be doing this anyway, because there are already large differences between individual intelligence. Whether those differences are racial doesn't really make a difference.

I think that's a pretty sober analysis - it's encouraging to see someone thread the thin line so publicly between populist ideologies on either side, with such a nuanced approach.

I would be more hesitant than him to call those like Nicholas Wade racist. People can be wrong without being malicious, and Jewish performance in the sciences has been impressive. You could definitely see how a perfectly non-racist person might see that and draw incorrect conclusions, even if you might disagree.

I particularly enjoyed this paragraph which I think highlights for me the point at which racism begins:

The real offense of racism, in the end, is to judge individuals by a supposed stereotype of their group - to ignore the fact that when applied to specific individuals stereotypes are almost misleading... Suppose you are the coach of a track-and-field team, and a young person walks on and asks to try out for the hundred-meter race, in which people of West African ancestry are statistically highly overrepresented... For a good coach, race is irrelevant. Testing the young person's sprinting speed is simple - take him or her out to the track to run against the stopwatch. Most situations are like this.

The Ethics of the Field

In the very last page Reich describes his ethical quandaries with his own work - he's grinding up the bones of ancient people who probably would rather their bones went undisturbed. Is that ok?

Ultimately he asked his uncle, the leading Open Orthodox Rabbi for his opinion. He gave an answer which is very in keeping with open orthodox philosophy:

...all human graves are sacrosanct, but there are mitigating circumstances that make it permissible to open graves as long as there is potential to promote understanding, to break down barriers between people.

Personally I prefer to look at this from a Kantian perspective:

We all have various things that we would rather not happen to us. Sometimes other people think our desires are silly. Now it's obvious from a utilitarian perspective that even if they are silly, they shouldn't act against are wishes if we'd find out about it because that will upset us. But what if they think we'll never find out?

If you're a Muslim or a Jew that could be someone slipping pig fat in your coffee, and if you're a vegetarian that could be, well, someone slipping pig fat in your coffee.

Would you rather live in the world where you get to disrespect the wishes of others in such cases, but in return other people get to disrespect your wishes, or in the world where everyone respects everyone else's desires, however silly, so long as they aren't exceptionally costly to comply with?

In other words, if we are prepared to dig up graves and grind up the bones, we have to accept that other people will one day do something we dislike equivalently to us, and decide that's a tradeoff worth making.

Put that way, I struggle to see digging up ancient bones as ethical. That's deeply frustrating, since I would love for this research to continue, but ultimately the ethical decision is hardly meant to always be the easy one.

Ancient Canaanites

With that the book ends, and I continues onto the next book - How to read the bible, by James Kugel.

Till the 12th century BCE there was a flourishing Canaanite culture in the Levant, mostly centered on the lowlands. After about 1150 BCE, the Canaanite culture weakened, and a much simpler Israelite culture developed in the highlands. Over the coming centuries that culture ultimately developed into the religion that would become modern day Judaism.

There are large cultural differences between the Israelite culture and the Canaanite culture, but there's also some interesting cultural connections. There's a long standing debate about whether the Israelites were descendants of the Canaanites who had fled to the highlands (possible following the late bronze age collapse), or nomads from the south who started leading a settled pastoral life.

Kugel leaves this as one of the questions which we'll possible never be able to answer. When I read that I thought that actually this is exactly the sort of question which full genome analysis could answer! All we need is some Canaanite DNA, some Israelite DNA, and some e.g. Moabite DNA, and compare the 3 to see how much of the Israelite DNA is Canaanite in origin vs Moabite.

Well Reich's book was published in 2018, and since then there's been a lot of advances in the field - including a study on Canaanite DNA. Whilst we're missing Israelite DNA, so we can't draw any firm conclusions, there's already some interesting results - for example, modern day Ashkenazi Jews are approximately 50% Canaanite (or at least Levantine) and most of the rest is European. Palestinians are even more strongly related.

This definitely suggests that the Israelite population wasn't completely nomadic, and incorporated a large percentage of the Canaanite population. It also makes the exodus much less likely since we would expect Egyptian slaves descended from Abraham to have either Egyptian or Mesopotamian DNA.

Finally it disputes both the Israeli claim that Palestinians are a recent invention who are mostly immigrants to Israel from at most a couple of hundred years ago, and the Palestinian claim that Israelis are European settlers with no connection to Palestine at all. Instead we're long lost relatives with strong connections to the land, and must learn to live together in mutual respect, and eventually, harmony.

I think this definitely serves to highlight the power of full genome sequencing - a question which has occupied dozens of scholars for hundreds of years with no real hope of ever reaching a conclusion, can be put to rest using standardized techniques. It seems likely it will have much the same effect on many other open questions in history.


History isn't just learning random facts about the past. It's about telling a story that people today can relate to and understand - it's the process of creating a map, not learning about the territory.

Comparing full genome analysis to carbon dating is apt. On their own neither can tell you a story - or at most a rather boring one. This happened before this. This group of people moved from here to here and mixed with that group of people. And yet they serve to provide the essential background on which the story can be written.

Without these techniques, two historians could look at exactly the same evidence and come up with whatever story they like. With them the stories are far more constrained. You can say whatever you like about relations between Neanderthals and humans, but somehow it's got to end in babies. In fact we can even put numbers on how often that happened.

So DNA analysis is likely to become the gridlines on which the vast map of history is painted. It might not be what stands out when you look at the map, but without it the roads would be all over the place and nobody would really know where they joined.

So should you read this book?

The book is certainly entertaining and informative. It contains lots of things that I didn't know across a wide range of topics, but at the same time I wouldn't say it contains any paradigm shifts that affect my daily life. If you're a student of history, archaeology or genetics I would definitely recommend giving this book a read. Otherwise read or skim if you think you'll find it interesting, and leave it if you don't.

New to LessWrong?

New Comment
4 comments, sorted by Click to highlight new comments since: Today at 3:02 PM

Whilst I was thinking of writing this book review for a while, I actually sat down and completed it thanks to the $500 bounty:

Thanks to Ruby for his feedback!

Thanks for the review!

Regarding the "countering racism" doubts, I can see how the results should disprove at least some racist worldviews. 

I think that an interpretation of human history among racists is the following: the population splits in to clusters, these clusters diverge in different "races", eventually one emerges as "the best" and out-competes or replaces all others, before splitting again. Historically, this view was used to justify aggressive expansionism, opposition to intermarriage, and opposition to any policy that could slow this process by helping races which were seen as lesser.

I think what he wants to say is that this picture is not supported by the genetic data, which shows instead population clusters which split and merge and split again among different lines, on a fairly fast timescale and without one population replacing the other (except arguably for the Neanderthals, but even then not completely). In other words, there's no darwinian selection at the racial level, and there has almost never been.

the movements of Polynesians across pacific Islands a couple of thousand years ago

In a nice coincidence, I read this review not a day after hearing about a genomic analysis study that did exactly this (published this September). It used the chromosome recombination idea to estimate when island populations split, and a different method comparing genetic variants to find which islands produced which new settler populations.

For those interested, here is the study, and here an article from Ars Technica summarizing it.

Oh! I hadn't seen you had published a review of this book too! It seems you explained much more in detail "The politics of DNA" than I did.

Something I regret a bit about my review of the book is not explaining better the different insights from America, India, Asia, etc. You tried to cover at least the most important points.

Cool book for sure! I am reading now: "The World Before Us: How Science is Revealing a New Story of Our Human" . I have just finished a couple of chapters, but I can already say that it is a good book too, with plenty of detail (like this one)