The Indo-European Advantage

In our book Greg and I devote a chapter to the hypothesis, developed with John Hawks and Doug Jones, that the expansion of Indo-European languages was driven by a biological change, an advantageous mutation, that enjoyed a large fitness advantage.  Mammals do not ordinary drink milk as adults, while all mammals drink milk as infants.  Lactase, the enzyme that digests the milk sugar lactose, ordinarily stops being produced after the age of weaning or so.  That is why your kitten loves to lap up milk while your adult cat, after a bowl of milk, may fail to reach the litter box in time.

Several versions of a broken regulator of lactase production are found in human populations leading to lactase persistence after childhood. Individuals with a broken regulator are called lactose tolerant (LT) because they are able to digest milk sugar. Without it milk may cause intestinal discomfort, diarrhea, or worse. On the other hand other people experience no discomfort at all: the difference is likely due to differences in the population of microorganisms in the gut. (There also exist other ways in which people are discomfited by milk, so the mapping between LT and discomfort is not clean).

There is a large region of homogeneity on European haplotypes with the mutation, telling us that it has arisen to high frequency within the last few thousand years. It must have enjoyed a huge selective advantage, but what was it? Seems unlikely that this prominent example of fast evolution was driven by flatulence.

Simple calorics seems to provide a simpler and better answer. In a dairy culture where fresh milk was readily available, children who could drink it obtained about 40% more calories from milk than children who were not LT.

Consider that 1 Liter of cow’s milk has

* 250 Cal from lactose
* 300 Cal from fat
* 170 Cal from protein

or 720 Calories per liter. But what if one is lactose intolerant? Then no matter whether or not flatulence occurs that person does not get the 250 Calories of lactose from the liter of milk, but only gets 470.

Many pastoralists consume a lot of milk (like the Herero I discussed in another post) but they must first ferment away the lactose to make the milk edible. It can be converted to yogurt or cheese or several other prepared milk products. This photo shows a Herero girl churning a gourd of fresh milk (along with way too much calf saliva and dead flies), converting it to something like buttermilk.


We also do that. Our 1 Liter of milk yields 100 g. of cheddar cheese, yielding 400 Cal of energy, cf. 720 in the fresh milk. We have discarded 300 or so Calories, about 40% of the energy, in the conversion.

Whenever food became tight LT children must had enjoyed a tremendous advantage over other children, an advantage that we can see in the genetic evidence of very rapid evolution of LT.

This entry was posted in Uncategorized. Bookmark the permalink.

34 Responses to The Indo-European Advantage

  1. Sean says:

    Calories are part of it, but all the empty calories in the world can’t compensate for protein deficiency, or nutritional rickets (cereal diet plus lack of calcium).

  2. Jason Malloy says:

    An NBER paper from several months ago applied a similar theory for the global power balance in 1500. Although it makes no reference to When Histories Collide.

  3. Matt says:

    “Whenever food became tight LT children must had enjoyed a tremendous advantage over other children, an advantage that we can see in the genetic evidence of very rapid evolution of LT.”

    Raw milk as kind of like a fallback food? And the LT advantage is more efficient processing as a fallback food?

    I’m sort of vaguely aware of a climatic change towards longer, colder and harsher winters with less productive summers for grain agriculture during a period of the European Neolithic which is associated with the time period of Indo-European expansion, combined with soil erosion.


    “The farming pioneers in Europe, though initially successful, eventually encountered problems which led to population crashes. Even some of their stock died out. Pig lineages associated with the LBK became extinct. Partly the problem stemmed from a shift to a more oceanic climate in northwest Europe, but the farmers themselves played a part in their own decline, triggering soil erosion by forest clearance.”

    If there get a sudden climate change catastrophe which places stress on milk as a fallback food, then that could explain why the generality of Neolithic Old Europeans could die off allowing Indo-European (or whatever lactase tolerant populations) could expand, without really needing any direct contact or warfare. If milk makes a good fallback food and the Old Europeans (the farming pioneers of Europe) had relatively low LT.

    • Matt says:

      “without really needing any direct contact or warfare”

      Although of course it’s still compatible with that.

      • harpend says:

        I didn’t have the idea of a “fallback food” in mind: pastoralists who depend on milk depend on a whole lot of it because herds do trash the environment for foragers and sometimes farmers.

        Herero ordinarily have so much milk that they even give it to dogs occasionally. But things got real tight over an eight year drought and children lost a low of body weight. There was nowhere near enough milk for them, especially considering that they were in direct competition with calves for the milk. Under these conditions LT kids would have survived, others maybe not.

  4. Ian says:

    Does the process which makes condensed milk affect the lactose content in any way?

    I ask because tinned condensed milk drinks, usually fortified with vitamins and flavoured, are popular with many Caribbeans. I would guess because most of their ancestors were from West Africa, that they’re lactose-intolerant as adults.

  5. Chris says:

    Obvious follow-up question: what is the selective advantage of lactose intolerance? It’s got to be strong, since it’s not hard for the regulator to break, yet we don’t see broken regulators in other mammals.

    • harpend says:

      No idea but the distribution does suggest that LT has a cost. LT is nearly fixed in northern Europe but gets uncommon around the Mediterranean: if it were cost-free then I would expect higher frequencies moving away from the Baltic.

      • John Harvey says:

        Scandanavia is relatively isolated geographically, so is it possible that the way of life, and hence genes, associated with a lactose tolerant population were able to spread here for longer without the genetic intrusion of neighbouring non-tolerant populations?

      • Greying Wanderer says:

        “if it were cost-free then I would expect higher frequencies moving away from the Baltic”

        Physical size? Height?

    • jb says:

      This is something I’ve wondered about too — in fact I was thinking about asking the same question, but you beat me to it. The body produces many digestive enzymes, and producing digestive enzymes involves only a small fraction of the body’s total energy budget (I would think anyway). How could this fraction of a fraction be a significant cost? Do we know of any other enzymes that are regulated this way — on during part of the human life span and off during others?

      • billswift says:

        The most important costs aren’t usually the costs of producing the proteins, it is the other effects of the proteins; the best known example being the “cost” of sickle-cell anemia in homozygous carriers of that gene.

    • Amandus says:

      Well, humans haven’t really had access to milk as adults to any greater extent untill quite recently (evolutionary speaking) when we embraced agriculture.

      Btw, how do you know that “it’s not hard for the regulator to break”? What about the appendix that doesn’t really serve us any function and can only cause us significant trouble when it breaks? We still have it because it isn’t that easy for evolution to just “get rid of things”.

  6. Sean says:

    You may be interested in this Low prevalence of lactase persistence in Neolithic South-West Europe Discussed by Maju at his blog here

    Milk is a Complete protein, it’s not just about calories because there is Kwashiorkor. It’s bit like meat and vitamin D. yes, ‘Meat consumption reduces the risk of nutritional rickets and osteomalacia” (Dunnigan, 2005). but as he says: “The mechanism by which meat reduces rachitic and osteomalacic risk is uncertain and appears independent of revised estimates of meat vitamin D content”.

    But there is obviously an big advantage to being able to digest lactose Actual lactase persistence more common than genes predict.

  7. harpend says:

    Is it not likely that there are other variants, SNPs or deletions or whatever, that also ding the regulation and hence are phenocopies of the common Eurasian SNP?

  8. Sean says:

    Insofar as I understand it, yes. As there are phenocopies doing somthing similar it suggests there must be very powerful selection for the trait. There is an interesting response to a comment though about genetic predictions of lactase persistence being more not being expressed in the actual phenotype frequency

  9. dearieme says:

    Don’t the Masai have a lactase persistence mutation that is different from the European one?

    • harpend says:

      There seem to be several in East Africa, all different from the main Eurasian SNP. There is another one on the Arabian peninsula and a different one, at low Frequency, in the Ukraine (as I recall).

  10. Bruce says:

    David Anthony’s “The Horse, the Wheel and Language” claimed that excessive milk consumption causes anemia and that the signs of anemia were detectable in the remains of children from the late 4th century B.C. Pontic steppe. Or so I remember.

  11. Ortu Kan says:

    One interesting consideration for equine-focused Inner Asian dairy consumption: mare’s milk has a higher lactose content than cow’s (6.2% as compared to 4.7%), though lower fat and protein (see here). It’s rarely consumed neat, of course …

    … but what’s rather remarkable is that some sources place the lactose content of koumiss in the neighborhood of 55 g/kg, exceeding the 46 of unfermented cow’s milk. I can’t say how typical these values are (I’d expect the former to vary with fermentation time and other particulars of handling) — and I don’t have numbers for horse cheese, dried curds, yogurt, etc. — but it’s something to consider.

    Lactase-deficient individuals’ generally greater toleration of yogurt relative to milk of equivalent lactose content suggests that some of these processing methods might work as much through directly seeding and/or enhancing the survival of certain bacteria in the gastrointestinal tract (increasing or prolonging microbial β-galactosidase activity) as through simple reduction of lactose content.

  12. David says:

    Re raw milk as a fallback food. It is perhaps significant that the Indo-Europeans became the dominant group in the northern parts of Europe where grass is a more dependable crop than cereal grains. It is also likely that Indo-Europeans had a more cattle-centric economy than that of the earlier neolithic settlers–a legacy of their history as pastoral nomads on the Eurasian steppes.

  13. Sonny says:

    Lactose tolerance beyond infancy is a neotenous feature. It seems to me most probable that it occurs in humans rather than other animals because of the amount of neoteny and longer childhood. Maybe it’s more likely to occur in environments where toddlers would more often have their survival rate depend on being able to be nursed because of usually available food being difficult to digest or seasonally scarce, where of course starting to be able to digest solid food would be helpful too (which is why there wouldn’t be a tendency to simply lengthening the time the digestive system is in its infantile state.) So I guess that genetic variations regarding lactose tolerance would be found, if enough data were available, to have been present in many regional populations long before domesticated animals were used very much for milking.

    (Current culture considers nursing of children past the age when they can walk and talk to be scandalous, but from the point of view of nature, it’s odd that humans out of all mammals are born not able to walk or even crawl the first day. Elephants have to learn to walk or else they can’t reach their mother’s milk.)

    As for Indo-Europeans being dominant in Northern Europe or anywhere else, it’s first of all a language family, only loosely speaking a people. I think every language group of the family probably had admixture with the language of some people who spoke a different language, and especially Germanic, which shifted its phonemes, stress pattern, and vocabulary much more than Hellenic or Baltic. (Celtic also shifted its phonemes in various languages more than Greek or Lithuanian.) The Baltic language Lithuanian was heavily relied on for the reconstruction of Proto-Indo-European, without which PIE would be more uncertain. Yet Lithuania is a relatively low area in Europe for lactose persistence, and the lowest area in Europe for the alleles that cause lactose persistence in Northwest Europe. (see which Sean linked above)

  14. Richard Jones says:

    Has anyone seen a response to William Durham’s argument in his book Coevolution about Indo-European myth and lactose tolerance?

    • harpend says:

      Thanks for the note: I am not familiar with the book. Can you hum a few bars of his argument? Sounds Interesting.

      • Richard Jones says:

        After analyzing the work on the subject by Lumsden and Wilson, Cavalli-Sforza and Feldman, Richard Alexander, Boyd and Richerson and others, Durham hypothesizes five categories and modes of gene-culture relationship:

        two interactive modes:
        1. genetic mediation (example: basic color terminology)
        2. cultural mediation (lactose tolerance and beliefs about drinking fresh milk)

        three comparative modes:
        3. enhancement (incest taboos)
        4. neutrality
        5. opposition (cannibalism and kuru)

        Inspired by the work of Bruce Lincoln and George Dumezil on IE mythology he focuses on the role that fresh milk drinking/lactose absorption plays in the various IE creation myths (Old Norse, Gaelic, Roman, Greek, Iranian, Indic) and how it varies nonrandomally with latitude.

        “for generations of Indo-European history three existed a latitudinal pattern of variation in cultural beliefs and values about milk. The traditional values coded in these ancient texts are of precisely the kind that would be required if culture mediated the genetic evolution of LA.” (p. 273)

  15. That Guy says:


    In reference to Maju blog post about Lactase Persistence in Northern Iberia, what he missed is that this is evidence of one of the earliest sites that showed skeletons with the LP allele.

    My take on it would be that R1b and cattle culture evolved near the Taurus mountains, as afterall Tauros (Latin) and Tarbh (Celtic) and similar words for Bull in Semitic point to a possible origin in the area. I think that there was a major range expansion down into Egypt and/or via Cyprus, to the Mediterranean Southern and/or islands to Iberia, there cattle herders spread North with R1b.

    R1 is found heavily North of the Caucasus, and probably entered the area from the North previously.

  16. Sean says:

    I don’t have the knowledge to argue about the genetic data stuff. But I’m afraid, for me, just looking at the at the denizens of West coast Ireland is pretty conclusive. That those folk are far less close to the original aboriginal European hunter gatherers than northern Scandinavians are, is not credible to me. For me, R1b being so high in West Coast Ireland notwithstanding, mtDNA is misleading about population movements.

  17. That Guy says:


    Can you rephrase, I’m not quite sure what you’re saying?

    What I’m theorizing is that R1b moved up the Atlantic facade, all the way to the Hebridies, and possibly beyond

  18. Sean says:

    R1b arrived in Europe in the Neolithic. OK, if R1b people replaced those originally inhabiting Ireland, the original inhabitants 0f Ireland would have been replaced by others, right?

    To me, the Irish have physical characteristics suggesting they are an older population, (ie they are are remarkably broad faced) . I think it is obvious that they are a survival from population that retained some of the look of the CroMagnons and are descended from the original inhabitants of Europe, on the maternal side at least. The mt DNA says otherwise, but I think that just shows mtDNA is misleading.

    • gcochran9 says:

      Sean, we have lots of autosomal DNA data, and the Irish are very close to the English.

    • pconroy says:


      I grew up in Ireland, my Father is Native Irish AFAIK, and my Mother is more Anglo-Irish, including a little Huguenot and Northern English, but when Doug McDonald analysed their results, My Mother mapped to around Bristol in England, and my Father to Plymouth, England.
      So both ended up looking like English people, and he even said that my mother was statistically 100% English.

  19. Sean says:

    “Looking like English people”

    Statistically I’m sure you’re right, and the Irish are doubtless closer to the British than other Europeans, but in regard to differences that make a difference one can look at appearance, behavour, and attainment. My great-grandmother was Irish, she looked very Irish. My fathers grandfather was Irish he looked like Barry Fitzgerald and had an incredibly thick head of hair. My grandfather (Irish /Scottish) was a very heavy drinker but never got a hangover, until a stroke at 48. His teeth met edge to edge, like a Cromagnon’s.

    The Irish have extremely high rates of schizophrenia HereA site says ‘Studies showed higher rates of schizophrenia among the Irish in the U.S. in 1913, 1920 and in the 1940s’. So whatever personality traits make people vulnerable to schizophrenia the Irish have more than other people. Schizophrenia is more common in the lower social classes.
    A fifth of people in Ireland are functionally illiterate – can’t read a medicine bottle.

  20. Rich Jones says:

    Have you read Ricardo Duchesne’s “The Uniqueness of Western Civilization”?

    Here is an interview with him:

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s