Skip

I am descended from Charlemagne!
May 11, 2002 9:08 PM   Subscribe

I am descended from Charlemagne! And you are too. I found tantalizing ideas in this Atlantic Interview of Steve Olson. Unfortunately, his Atlantic article is not available (for free, anyways). He mentioned, in the interview, the work of Humphrys and Chang. A fwe Google searches later, among a labyrinth of pages about Royal descents, I FOUND! what I was looking for [More inside]
posted by vacapinta (13 comments total) 1 user marked this as a favorite

 
Humphrys is a geneaologist. Chang is a statistician who has constructed a mathematical model of ancestry.

If we define the Most Recent Common Ancestor (MRCA) as the most recent person in human history that is a common ancestor of *everyone* alive today then

Chang's model with Humphry's assumptions predicts:

- Our MRCA may be as recent as someone living in 1200 AD.

- Humphrys makes a strong case that Western Europe, with its strong mixing, may have a very recent MRCA.

- Before 700 AD, every single human is either ancestor of no one alive today, or ancestor of everyone alive today. So the Islamic Muhammad, the Irish/Celtic Niall of the Nine Hostages, the English/Saxon Cerdic, and the Continental/pre-Norman Charlemagne, are all ancestors of everyone alive today. In fact, the model predicts that 80 percent of the entire population at this time is an ancestor of everyone alive today.

- Between 700 AD and 1200 AD, every single human is either an ancestor of no one alive today, ancestor of everyone alive today, or ancestor of some people alive today.

- After 1200 AD, every single human is either ancestor of no one alive today, or ancestor of some people alive today.

Gladiator tidbit: About 80% of any group of Romans you might see depicted in history are your ancestors.
posted by vacapinta at 9:10 PM on May 11, 2002


This came up recently in a course I teach -- as a result of class discussions, I put up a page with some further information and plots from some new simulations. The basic result is very robust, and continues to apply (at somewhat larger time scales) even if most mating is local and there is very little exogamy.
posted by myl at 9:37 PM on May 11, 2002


So how does this work if you're a Native American or Chinese?
posted by Poagao at 9:53 PM on May 11, 2002


How long before everyone on earth is related to me? (Or you.)
posted by 23skidoo at 10:44 PM on May 11, 2002


damn, there goes my hobby....
posted by geneablogy at 10:54 PM on May 11, 2002


and then, recent human genome mapping indicates that all humans are essentially brothers, at least on the genetic level...
posted by bwg at 3:59 AM on May 12, 2002


and then, recent human genome mapping indicates that all humans are essentially brothers, at least on the genetic level...

The difference is that the "most recent common ancestor(s)" uncovered by genetic comparisons (e.g. "mitochondrial Eve") lived about 200,000 years ago. By contrast, Chang and others posit "most recent common ancestors" that are a couple of orders of magnitude more recent, roughly a thousand years ago or so.

This difference is why biologists tend to be skeptical and even upset by Chang's results. But it arises because the data and models used in genetic comparisons consider only one-parent descent (e.g. in the maternal line for "mitochondrial Eve") while Chang models two-parent descent.
posted by myl at 4:31 AM on May 12, 2002


How long before everyone on earth is related to me? (Or you.)

According to Chang's theorem: with about 80% probability, and assuming random mating and constant population size, everyone will be descended from you after about 1.77 * log2(N) generations, where N is the population size. With 20% probability, no one will be descended from you.

I believe that there are now about 6.2 billion humans on earth. This yields an estimate of about 58 generations (if you're one of the 8 in 10 who are universal ancestors). This is the expected value, but the variance of the process is amazingly low. If there are roughly 4 generations per century, this would be 1,450 years. At an average of 5 generations per century, it would be 1,160 years.

Adjusting the assumptions in realistic directions (mostly-local mating, increasing population, etc.) increases the time somewhat without changing the final result.
posted by myl at 4:45 AM on May 12, 2002


So how does this work if you're a Native American or Chinese?

It may not change things much.

Consider the following thought experiment:

There is a hidden region full of people that have been (completely) genetically isolated from the rest of the world for many thousands of years (probably true for the Americas a thousand years ago, but never true for China). At a certain point in time, one single person from the outside joins this group as a parent. With 80% probability, after a few tens of generations, this outside individual will be a universal ancestor for the isolated group -- and so will all his or her ancestors.

It doesn't take very many "intermarriages" -- as few as one! -- to set up this situation. I suspect that the rate of intermarriage through central Asia over the millenia has been great enough to ensure that modern Chinese are descended from ancient Greeks, and that modern Greeks are descended from ancient Chinese (for example).

It's less clear whether the 500 years since Columbus, or the longer time since the Norse colonies, have been enough to ensure that modern Native Americans are all descended from El Cid, or that modern Spaniards are all descended from Montezuma. My own guess (for what little it is worth) is that the answer would be different for different subpopulations, but overall it would be "no".
posted by myl at 5:00 AM on May 12, 2002 [2 favorites]


Is he positing that the rate of gene flow between populations has remained constant through history, or that any change in rate will have little effect? It seems that he is proposing this one equation as being true for all points in time.
This is really cool, I'm just having trouble getting my head around it.
posted by nprigoda at 6:08 AM on May 12, 2002


Thank you, myl. I wish I had found your page before posting.

This is strictly about genealogy and not genetics. I may share no DNA with my great-grandfather but he is still my great-grandfather.

Its really about the power of geometric progressions. Imagine a man from China fathers a child with a woman from Latin America. The child inherits the genealogy of two continents! If that child's children then go off and have children with Western Europeans and Africans etc. its not hard to see intuitively how quickly everyone's ancestry gets all mixed up together.

It sounds dramatic to say that Charlemagne is my ancestor (I am mostly Spanish with French and native Mexican thrown in) but the truth is of course that a big chunk of the people alive then were also my ancestors including many peasants and criminals. :)
posted by vacapinta at 1:17 PM on May 12, 2002


Geneaology is such a keen thing. The standing joke in my family is if that a carefully chosen twenty thousand or so people die, my oldest brother will be King of England.

So you guys watch out. :)
posted by ebarker at 3:52 PM on May 12, 2002


Have you considered this bump on your road to the throne?
posted by dmo at 11:16 AM on May 27, 2002


« Older Maybe NISSAN fights for domain names a little too...   |   Saudi Arabia considers... Newer »


This thread has been archived and is closed to new comments



Post