Here is a puzzle for you to figure out: did Genghis Khan really have 16 million direct male descendants? Note the careful wording: direct male descendants. It is a factoid that has been around since 2003 when a now famous genetic study concluded that 16 milllion men in Central Asia share a remarkably similar Y-chromosome, meaning that at some time in the past they had the same forefather. The factoid that it was Genghis Khan has been doing the rounds in the international media ever since.
How can they know such things? Well, because they know how many copying mistakes there are in each generation in bits of the genome, so by looking at how much variation there is in a population they can work out how long ago they started with the same chromosomes. With a bit of statistical trickery you can also find clusters of men who will have a later common ancestor, essentially by first divvying up a large population in more homogenous ones that have similar Y-chromosomes.
Now, it is true of all men that if you go back far enough that they will have the same forefather, but the unusual things about the 16 million in Central Asia is that their common forefather appears to have arisen around the year 1000, give or take a century (and with Genghis Khan you have to give 1.5 centuries, which is your first main clue). The year 1000 is rather late for so many related people however.
The factoid is thus that Genghis Khan, who had many sons and whose many sons had a lot of sons themselves currently has 16 million male descendants. This is 8% of the Asian men studied in the 2003 study, but more than 8% in particular areas. With a roughly tenfold general population increase in that region, Genghis would have been 1.6 million times as successful as any other man living in that era.
I encourage you in the comment thread below to say if you believe this and if so, why. If not, tell me why this factoid, which to my knowledge has been unchallenged till 2012 would in fact not be true.
I will let you know on Monday what I think about it.
As a bit of background on those who believe that because something has been unchallenged for a while that it must be true: the world is full of factoids that survive in the popular media and even in the scientific community for a while, even if on closer inspection they are not all that plausible. An obvious example of a now ‘disproven factoid’ is whether the Great Wall be seen from outer space or the moon? It seemed plausible because it is so long. However on reflection you should already have your doubts: the Wall is less broad than your average house. It is basically as visible as a very long lane with large trees of a slightly different colour to the rest. Now, if you can’t see something as thin as the average house or a lane of trees from outer space, why would you be able to see a whole string of things smaller than the house or a large tree? So with a bit of common sense you could already have known it was unlikely that you’d be able to see the Great Wall from outer space (unless with special equipment, of course, and in closer orbit you might get a glimpse). It took a few decades but ultimately, indeed, we found out you can’t see the Great Wall from the moon and that what was bandied around for decades was a myth.
Perhaps the same is true of this factoid also?
I just like the idea that my wife is the several times great grand daughter of a Mongol emperor.
I’m mostly Irish, with the (ostensibly) Viking red hair gene, so we have a reasonable chance of producing offspring that are Ming the Merciless.
sancho maybe you are really a Varangian
I don’t think I could ever work for a Basil, but interesting wiki page.
Autocrats often can’t rely on their underlings for loyalty, so they employ foreigners. That’s where the money will be for modern mercenary companies like Xe when the US wars dry up.
Romantic notions of oath-bound, loyal-unto-death fighting orders are a bit, well romantic, though. Feudal knights mostly had the nobility of muggers, and as Robert Evans wrote, taking the samurai codes at their face value “is the equivalent of reading a high school handbook and determining that teenagers live by a strict code of attending class and turning weed dealers in to the cops.”
“that teenagers live by a strict code of attending class and turning weed dealers in to the cops.”
We were pretty good at the image of well behaved.
Byzantium spent a lot of time and money on playing the numerous step tribes off against each other ,keeping them too busy to trouble the Danube frontier.
There is a letter to a emperor from one aggrieved local steps chieftain complaining about his worthless lying neighbors that had been awarded the ultimate inducement ; they had been allowed to move to Greece where they would unjustly enjoy ” wine, olive oil, lemons and hot baths ”
Byzantium was also a bureaucratic state, whether the autocrat was Porphyrogénn?tos, or a bastard made little difference.
Indoor plumbing’s pretty good. I’m glad I don’t have to kill any of my relatives and move to Greece.
I don’t think it is terribly unbelievable (i.e., I would want real evidence, but I wouldn’t dismiss it). Let’s say Ghengis had 100 children (certainly plausible). If each of these had 10 children each then it only take 100 * 10 * 10 * 10 * 10 * 10 five extra generations to get there (say 150 years) excluding interbreeding etc .. Even if, apart from Ghengis, they wern’t very successful (let’s say each child had just 3 children each), then you only need about 10 generations to get there (100 * 3^10 = about 6 million).
If I haven’t messed this up then, let’s assume about 30 generations have gone by, which is conservative given people would have been having kids at 16, then you get 10000000 = 100 * n^30 = Each of his descendents must have had 1.46 kids each.
Started doing the maths, but the “direct male descendants” bit makes it tricker. Can’t see why not, though. We are talking about perhaps 32 generations. If each male descendant had two male descendants of their own, that would give you 4.3 trillion direct male descendants without overlaps. So 16 million doesn’t seem much of a stretch. (The same sort of calculation can be used to show that almost every Caucasian in Europe is descended from Charlemagne.)
The calculation would be sensitive to the number of kids each descendant had. There’s probably an opportunity for a Monte Carlo simulation in there somewhere.
Is there a university near where you work? See if they have a maths department and get someone there to help you.
Quick! To the Bat Mitochondria!
We do not have Genghis Khans actual genome.
The estimate of the rate of variation of the chromosome over time is a estimate, therefore the ‘common’ ancestor might have been sometime before or after Genghis Khan.
I would also guess that there is a good chance that the ‘common ancestor’ referred to might be in reality a group of uncles and other male relatives in a close knit tribal sub-group of about 1000 people that later came to dominance.
Conrad
No idea what the infant mortality rate was like but I bet it was high, you would need to produce a awful lot of babies to get 100 living long enough to breading.
If you want just male descendents (and using the correct numbers rather than just being slack) you need something like: 16,000,000 * 2 = 50 * (n/2)^30 = 3.12 kids per male. I use the * 2 to because you want twice the number at the bottom level since half will be female and 50 males to start with. You can use n/2 because that wipes out the female chunk of the tree caused by females.
John, if you want infant mortality (and infertility etc.), you can just add whatever percentage you want onto the previous number. So, for example, if there was a 50% casualty rate, then you’d obviously need to double the number (i.e., 6.24 kids for males that have them). This still doesn’t seem exceptionally unreasonable.
The study is based on a surrey of 1437 men from 18 ethnic groups , that is 79 individuals per group , not that big a sample size per group . The study also found little evidence of the gene in other groups in areas that were occupied by the mongols for centuries – dosn’t sound right.
Actually, thinking about that, it’s an overestimation, because the 32 million includes those that are still alive but not in the final generation.
Well I certainly remember when I was a kid my Viennese relatives occasionally saying “Don’t mention Great Uncle Gengis”. Come to think of it the other side of the family used to say the same thing.
Are you sure it wasn’t an actual uncle that would turn up, drink a litre of slivovitz and rail against Jews?
Well if he was a Viennese uncle of mine it might be hard to explain why he’d rail against Jews.
Your actual question is what sort of assumptions would you need to make so that one man, who was 1/1.6million living Asian men, can have cumulative genetic success so that his current living direct male descendants are 8% of living Asian men 850 years later.
The following scenario achieves this in less than 200 years
1. Khan has 50 sons reach adulthood. No unreasonable to assume he has more than 100 live birth sons (50% infant mortality) given multiple decades of copulation with any woman he fancied.
2. Half of Khan’s sons have 25 sons reach adulthood, the other half have 5. This is to restrict the second generation dominance to only those sons that gained from his conquering shenanigans. Others he leaves behind at villages breed a lot because men are in short supply for a while after one of Khan’s murderous raids. We are already up to 750 male descendents during Khan’s lifetime.
3. Khan’s grandsons and the next 3 generations have 5 sons reach adulthood. We might assume that rebuilding populations after war might have meant high population growth over this period, dominated by the sons Khan and his sons leave behind.
4. During the century after Khan’s birth the Mongols kill 10% of unrelated males in Asia (say 200,000). Some variation on this idea of eliminating the genetic competition helps the story.
Khan’s male descendants are already around 8% of the Asian male population, and that’s only 5 generations. The logic of the folk story rests of the critical assumption of how other male lineages are cut off through war and social dominance of Khan’s descendants to generate cumulative genetic dominance.
I find it plausible. Some of those eastern emperors had thousands of concubines. They were on the payroll and their offspring had superior status and therefore the sons had privileged access to women. So 100 sons is not much. I have no idea how many official women Khan had or for how many generations his official offspring inherited privileged status.
The report of the study is a bit short on detail , it states that 1437 men from 18 ethnic groups were sampled – I assume that is about 80 individuals per ethnic group were sampled. The report states that 35% of the 80? mongols sampled carried the gene – i.e about 28 individuals. Is this likely to be a statistically reliable result?
The reporting of the actual study is causal “direct male descendants” should really have read male descendants of the small group of males that Genghis was the leader of.
I think pauls interest in this research provably comes down to this:
The problem with all the exponential extrapolations is that they assume no-one marries cousins, whether 1st, 2nd…nth. That’s a pretty absurd assumption to make. The Mongols may not (or may) have been as incestuous as the European nobility, but there’s limits on anyone’s partner pool. This is even more so when you still live in villages, as have most people until very recently. You can’t have a child with someone unless you meet them physically (at least before sperm banks), and the odds of meeting someone in a village context before decent transport to whom you are not related in the past 5 or so generations (let alone 30) is very very slim.
Look at the other direction. If I assume no interlinkages in my family tree, at any distance relationship, I get roughly 8589934592 ancestors over 1000 years – more than the population of the Earth as it is, despite the fact it is now exponentially more populous than just a few generations ago.
“The problem with all the exponential extrapolations is that they assume no-one marries cousins, whether 1st, 2nd…nth.”
I don’t see why inter-linkages really matter apart from your assumption that people don’t like too much of it (which at least for cousins, is perhaps not too good an assumption). Ignoring genetic problems, for the sake of this argument, it doesn’t really matter if someone breeds with their sister, as long as they produce enough males.
It is easy to get confused with these kinds of calculations. Remember Gengis was born in 1162 which is 38 generations of 25 years ago. How many great^38 grandparents do you have? Well 2^38 is about 208 billion. Doesn’t make sense. The problem is with the whole idea of a tree. Yes – everyone has a tree but my tree intersects with yours and the further back we go the more overlap there could be. The tree image is very misleading.
You claim he was 1.6 million times more successful than everyone else because the population increasing by a factor of 10 and Gengis had 16 million progeny. It is quite possible mathematically for everybody wo lived at the same time as Gengis to have 16 million descendants, even with a stable population. They woudl just be the same progeny. We are all cousins.
Statistically it is feasible that there could be 16 million male descendants of Genghis, however whether it is factually correct we’ll never know.
We don’t have Genghis’s DNA in order to identify his offspring from other individuals of the time. We don’t know how many children he had (from his harem of 2000-3000 wives).
We cannot even confirm which of his offspring were really his children (even at the time Jochi’s parenthood was questioned), or whether their children were really their children.
What we do have is a belief that a particular genetic line has been more successful than others and have assumed that indicated an origin from the ‘top dog’ human in the time period.
That may not necessarily be true. In many societies – human and other animal – the top dog isn’t necessarily the most successful breeder.
So it is a nice theory and many people might feel uplifted by thinking there’s a little Genghis in them.
However all I know for sure is that Bill Prosser is a direct patrilinear descendant of Genghis, as Douglas Adams says so.
“…many people might feel uplifted by thinking there’s a little Genghis in them.” I’d rather some Casanova of Newton. I guess I’m shit out of luck with the latter,