Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
You (maybe?) have thousands of ancestors from the 1600s (basqu.es)
36 points by kaycebasques on Sept 5, 2023 | hide | past | favorite | 71 comments


If you actually go back and look at who those people are, though, you likely don't have anything like as many as calculated, because your family isn't so much a tree as a DAG—there's plenty of duplication.

For example, my wife and I recently discovered that we're sixth cousins. This means that when our kids trace back their ancestry, there's an entire subtree on my side that is duplicated on her side, which dramatically shrinks the number of ancestors down from the theoretical maximum. And this is far from an uncommon occurrence—we most likely have many other subtrees in common that we simply don't know about.


My parents are related through one set of grandparents four or five generations back.

My wife and I are related through the same grandparents.

These are weird, but it makes sense - we're from smaller towns, which were settled not that long ago in the grand scheme of things. If you trace enough of us back, we're all probably related.

But the seriously trippy part is - my younger sibling and their spouse are related through the same great(x4 or so) grandparents. The sibling is from a different country that isn't even close to where we're from. Like literally the other side of the planet.

It blew my mind.


If one of you is from one corner of the former British empire and the other from another corner, it becomes a lot less surprising. The British showed out everywhere the past ~400 years and often forcibly relocated certain subjects (e.g. the Irish).


[flagged]


Don't be disingenuous and provide the end of the quote:

> four or five generations back.


I think it was just ambiguous wording. “Through the same grandparents [as mentioned earlier]” vs. “through the same grandparents [that we directly share]”.


My old man, bless him, found one instance of inbreeding going back to 1800 for each one and going back to 1700 on a few.

My son's mother's ancestry is very different from mine.

If anything I think we might live in period of the greatest number of long distance gene flow events in history. And I think this has been true for hundreds of years at this point.

EDIT: That's not to say inbreeding doesn't hugely reduce your number of ancestors, just that the modern world for many social groups has much less inbreeding than earlier societies.


OK, so this is the key assumption that does not hold up in the real world?

> With fully random matings the number of your ancestors doubles every generation.


No, because fully random would still cause duplication, because the population of Earth is finite not infinite. With matings specifically selected to create no common ancestors back to X generations, then you theoretically have 2^X ancestors.


I'm kinda grokking it but still have no clue how far off the calculation is. To go back to the blog title, are you saying that the number of your ancestors is more like hundreds rather than thousands? Or is it just impossible to even put a number behind it due to how complicated the real-world relationships are?

Edit: Just saw your other comment here. The answer seems to be more along the lines of "impossible to put a number behind": https://news.ycombinator.com/item?id=37387655


The latter. It's impossible to make any generalizable statement about everyone's ancestry, you really have to look at the family tree and count them up one by one.

People whose families have been in one place for generations are much more likely to have a smaller set of ancestors than people whose families migrated many times over the last few centuries. And any given person's family tree is likely to have some subtrees that were sedentary and others that migrated, further complicating any attempt at an abstract calculation.


Fully random only appreciably diverges from 2^n when 2^n becomes a sizable share of the world's population.

When the population is roughly 2^2n, you expect to lose one ancestor in that generation.


It absolutely does not. Going backwards in time like this, most people will have many generations where the number of their ancestors in that generation go down, not up.

For the majority of the past four thousand years, almost everyone lived in farming villages that had ~50-100 families living in them, and mostly married inside those communities. Even just going to a neighboring village for a bride was something that only happened occasionally, but even when that happened, it was almost always the same few neighboring villages. This means that when you trace ancestry backwards, you always end up related to everyone living in a few villages, and then every few generations this extends to a new village. Then if you are related to a successful trader or soldier or something, it will suddenly double the amount of ancestors in a generation, but I really need to stress that this is rare.

The introduction of the bicycle did a lot for reducing inbreeding in farming communities, because it was something that a normal people could own, and allowed day visits to neighboring communities. Before that, it was entirely ordinary to live and die without ever traveling more than 10 miles from the place where you were born.


I guess it doesn't, or you could go back to the time where the early human lived expecting an astonishingly high number too, or even before the big bang if you wanted.


We are all basically sixth cousins from what I've seen written.


As long as you're comparing two people from the same country and ethnic group (which tend to be the most common pairings for childbearing), it's pretty close to true. It doesn't work so well if you're comparing people on different continents.


A country is too big. The number of 6th cousins somebody has is on the order of thousands. If you're from a small town with stable population (people haven't moved in or out much for the last couple hundred years) you might be 6th cousins with pretty much everyone.


many earlier societies had traditions of inter-tribal wife-stealing .. farmer cultures with socially important grandmothers practiced social interchange.. meanwhile, opportunistic breeders often physically moved around and did not have stability patterns.. this is a massive subject, not possible to talk about without errant conclusions and misstated sub-parts, right?


None of that changes the math of how many cousins you have.


Agreed, sixth is too close to be "everyone" in a country, but the general principle is accurate: we tend to have closer common ancestors than we might think.


> With fully random matings the number of your ancestors doubles every generation. Ten generations ago (250-300 years) we might expect about 1,000 ancestors each contributing 0.1 percent of their total variants to your genome... when the number of ancestors exceeds 2,000-3,000 (12 generations would be about 4,000), it is likely that some ancestral sections of DNA have been lost altogether... by about 15 generations back many of the ancestors contribute no DNA variants at all but they are still ancestors on the lineage.

So cool, I never thought about a generation limit to meaningful genetic contribution.


I guess from a genetic standpoint, it doesn't really matter. Since many of those near-relations (as other comments have mentioned) who mated and had children would have somewhat similar genetic profiles, one particular SNP from a great-x10 grandparent might still propagate to us, simply because other mates would also have that gene.


Multiple people intuitively pointed out that the OP is actually quite wrong. The shocking fact, to me at least, is how far this goes: all of humanity share a single male ancestor [1] (and no, it was not the first man to ever live), and a single female ancestor [2], who likely lived thousands of years apart.

[1] https://en.wikipedia.org/wiki/Y-chromosomal_Adam [2] https://en.wikipedia.org/wiki/Mitochondrial_Eve

PS: see also references fromthe linked wikipedia pages


The genes book I mention in the post talks about Y-Chromosomal Adam and Mitochondrial Eve. Super cool stuff. The directed acyclic graph nature of lineage totally makes sense why the number of ancestors is smaller in reality than the theoretical numbers. I don't follow how Adam and Eve relate, though. I know this example is crude but bear with me; hopefully it's valid enough to have discussion around. If we follow the lineages back to the point where we have 1000 ancestors each, it seems like Adam just represents the 1 we all have in common, and by extension the other 999 are the ones not in common.

From the Wikipedia article:

> Although the informal name "Y-chromosomal Adam" is a reference to the biblical Adam, this should not be misconstrued as implying that the bearer of the chromosome was the only human male alive during his time.[7] His other male contemporaries may also have descendants alive today, but not, by definition, through solely patrilineal descent; in other words, none of them have an unbroken male line of descendants (son's son's son's … son) connecting them to currently living people.

(Reading the Eve article left me with even more questions than answers, but I'm still getting the sense that Adam and Eve are orthogonal to the "number of ancestors in 1600" topic.)


This makes some crazy assumptions that not only do people not have kids with their first cousins (they do!) but they don't with there Nth cousins for N going into the dozens, with the family tree organizing itself into a perfect inverted pyramid. Also a generation can be 15-50 years, and is unique to each parent-child pair in the DAG (to use the term from the sister comment).


I'm kinda wrapping my head around the DAG thing (but still have no clue how large of an effect it has on the overall calculation) but don't get how 25 years per generation is off. I know that in the last few centuries people have been having children at older and older ages but you're saying that an average like that doesn't hold for the long-term? Or maybe it's just very complicated? E.g. some 60 year-old man gets a 20 year-old woman pregnant 700 years ago and it has a dramatic change on the calculation?


Historically, having your first child around 27 was average except for the wealthy. The post war period was a historical aberration of young marriage.

> fathers consistently older (30.7 years) than mothers (23.2 years)

https://www.science.org/doi/10.1126/sciadv.abm7047


Those numbers from the abstract are average generation times, which is very different than the average age of having the first child.


The problem isn't so much the average generation length but that generations are not coordinated waves each only having kids with the same generation.


Try the awesome numberphile video "EVERY baby is a ROYAL baby" [1] https://www.youtube.com/watch?v=Fm0hOex4psA


It’s very strange to look at a long genealogy of oneself and be reminded that you only know the last names of your grandparents and maybe great grandparents, but that 300 years ago, almost all your descendants would’ve had totally different names. Everyone wants to know who their father^N was, just because we’re attached to a silly little name.


Wouldn't that be ascendant? Or well ancestor to use the common term.


Yes my mistake. Progenitors I guess would work to.


Because of cousin marriages the number of ancestors is reduced, it's called Pedigree Collapse. In certain populations it's very evident if anyone tries to build a genealogical tree with the help of a dna test.


Thanks for the lead. Someone in another comment described it as an inverted pyramid. The graph of Cleopatra VII's lineage seems to be the best example of that: https://en.m.wikipedia.org/wiki/Pedigree_collapse


Fractal Patterns in Reasoning - David Atkinson, Jeanne Peijnenburg - 2012

https://projecteuclid.org/journals/notre-dame-journal-of-for...

> This paper is the third and final one in a sequence of three. All three papers emphasize that a proposition can be justified by an infinite regress, on condition that epistemic justification is interpreted probabilistically. The first two papers showed this for one-dimensional chains and for one-dimensional loops of propositions, each proposition being justified probabilistically by its precursor. In the present paper we consider the more complicated case of two-dimensional nets, where each "child" proposition is probabilistically justified by two "parent" propositions. Surprisingly, it turns out that probabilistic justification in two dimensions takes on the form of Mandelbrot's iteration. Like so many patterns in nature, probabilistic reasoning might in the end be fractal in character.


FWIW, a free site for finding and preserving info about your ancestors, near and distant, is at https://www.familysearch.org . I used to work there; they have put a lot of resources into making this available and preserving the data long-term, permitting collaboration with others who have overlapping lines, etc.

I believe it also has a feature somewhere for determining how closely related you are to someone. My wife and I discovered we are 13th cousins (or so).

And it has a really broad wiki to help you do accurate genealogy, with many country-specific tips and other pointers to other resources.

It will not disclose info on living persons, in the tree, except to that person (and maybe their children? I forget).

Lately I have not been active on the site and get less than one email per month from them, I think.


Lots of those nodes are not unique - especially if you're from breeding ground with more constraints (e.g. Iceland).


An implication is that any famous person from long enough ago, if they're not totally mythical and had any children at all, is your ancestor.


It's not enough that they "had any children". For example, if they had exactly one child, and that child was childless, then they are not your ancestor.

Biologists define the "identical ancestor point" to be the most recent time X in the past such that every person alive at time X is either an ancestor of everyone alive today, or of no-one alive today. Computer simulations suggest that the identical ancestor point is quite recent in geological terms, e.g., it's probably within the past few thousand years (it would be even more recent if we ignored remote places like Tanzania). See: Rohde et al (2004), "Modelling the recent common ancestry of all living humans", Nature. https://www.math.arizona.edu/~bhallmark/TMRCA/naturerohde.pd...


I assume you mean Tasmania here.

From the linked paper:

> In the case of Tasmania, which may have been completely isolated from mainland Australia between the flooding of the Bass Strait, 9,000–12,000 years ago, and the European colonization of the island, starting in 1803 (ref. 13), the IA date for all living humans must fall before the start of isolation. However, the MRCA date would be unaffected, because today there are no remaining native Tasmanians without some European or mainland Australian ancestry.


Truganini may passed but members of the Pintupi Nine and other groups with no European blood remain alive.

Oddly, portions of mainland Australia may have been as genetically isolated from each other as Tasmania, once particular environments were settled (ocean, river, desert, etc) they appear to have remained stable.

https://en.wikipedia.org/wiki/Truganini

day the Pintupi Nine entered the modern world https://www.bbc.com/news/magazine-30500591

Aboriginal mitogenomes reveal 50,000 years of regionalism in Australia https://www.nature.com/articles/nature21416


I’ve been thinking about this a lot.

Each one of those people had a name (and a life, and thousands of ancestors themselves).

How many names of these 32,000 do you know?

While for many of my lines, I know the names to the 1600s and further.

Other lines dead end in the mid-1800s…what stories am I missing because of these all these dead ends?


> How many names of these 32,000 do you know?

Enough to bore my family! Genealogy is fun, at least for me...

With modern technology, it's easier than ever to dig into the primary sources and find out more about your ancestors.

But I would warn everyone to stay away from the big commercial genealogy companies. For one, they have extremely bad privacy practices (they will never, ever, stop spamming you), for the other, they have no interest in the truth and will happily tell you what you want to hear to keep you as a customer (they encourage very bad genealogical practices).


This ^

I use them for pointers to resources.

Archive.org is my primary research tool.


>I know the names to the 1600s and further

That is, assuming the officially recorded parents are the actual biological parents. It's not nice to talk about in polite society, but sometimes people are mistaken about biological parenthood, most often fatherhood but sometimes motherhood too (think: babies swapped at the hospital). All it takes is one error to invalidate the whole line behind it (though, eventually the line will hit your ancestors again anyway).


Yes, trust but verify. In my lines, the mother dying in childbirth is a much more likely scenario.


How many ancestors up to the 1600s do you know of in your case? If you count up every known ancestor up to 1699 (or 1600), what's the total?


2,096


This is why the word "lineage" doesn't make much sense. Go back far enough and everyone is your ancestor. Assuming you have enough progeny, then go forward far enough and everyone is your descendent. It's more like a giant braid than a giant tree.


> Go back far enough and everyone is your ancestor.

That's incomplete. It should be go back far enough and everyone is either your ancestor or is nobody's ancestor.


The more that I think about this the weirder it gets. What fundamental error am I making here? How can I have 1K ancestors 10 generations ago, 32K 15 generations ago, 1M 20 generations ago?? But when I start with my parents, and think about how they have two parents each, and then go from there, the numbers seem to check out...


You do not necessarily have 4 grandparents. Try to think of ways in which you can have fewer than 4 (hint: it's a little bit impolite to talk about). Neither do you necessarily have 8 great-grandparents, for the same reason.


We're talking incest? And the theory (or fact) is that it was a very widespread practice throughout human history? Trying to wrap my mind around how that changes the calculation... it would drop the number from 2x every generation down to what?


We only call it incest when the nearest common ancestor is a parent or (depending on the jurisdiction) a grandparent.

If your nearest common ancestor is a great-great-great-great-grandparent, that's not incest (it would most likely not even be known at the time of marriage), but it still leads to large amounts of duplication in your family tree. Every family tree will have many instances of nth-cousin marriages simply because people tend to marry people who live close to them.

It's very hard to calculate what the actual number would be because it depends a lot on where your ancestors lived and how much they migrated over time.


"Inbreeding" was the word I was looking for, methinks


> We're talking incest? And the theory (or fact) is that it was a very widespread practice throughout human history?

That really depends what you're imagining when you say "incest". In literal terms, yes, we're talking about mating between two people who share common ancestors. Those do not need to be particularly recent.

In genetics there is the concept of the "coefficient of inbreeding" which tracks how much of an individual's genome is identical-by-descent across both chromosomes. (That is, if you inherit a particular stretch of DNA from your mother, and you inherit the same stretch of DNA from your father, and the reason they are the same is that your father and your mother both inherited that DNA from the same person, then that stretch of DNA contributes to your coefficient of inbreeding.)

If you're imagining brother-sister marriages, that is a much higher level of inbreeding than would be common in human history. (Expected coefficient of inbreeding in the offspring: 25% assuming zero inbreeding in the parents, perhaps because they are themselves the offspring of a cross between an African Pygmy and an Australian Aborigine.)

But it is true that different populations have more or less inbreeding than others; we can say that Arabs tend to be more inbred than the French. This is not necessarily because they tolerate incest at closer degrees of kinship than the French. It is mostly because they have a greater quantity of still-relatively-remote incest. (Not the greatest example; Western Europe hugely stigmatized cousin marriage under the influence of the Catholic Church, so in reality Arabs do tolerate closer degrees of kinship in a marriage than the French do. But the concept is sound - inbreeding will "build up" in a population over time.)



> What fundamental error am I making here? How can I have 1K ancestors 10 generations ago, 32K 15 generations ago, 1M 20 generations ago??

You can't. The population of the world gets smaller as you go back in time.

There are that many open spots in your family tree, but, as you go back in time, you will see that each person in the tree simultaneously fills a large number of different spots.

A related but not identical observation: the set of all your ancestors is roughly 1/3 male and 2/3 female, despite the fact that there are an equal number of male and female spots in your genealogy. Males are doing more double duty than females.


> the set of all your ancestors is roughly 1/3 male and 2/3 female, despite the fact that there are an equal number of male and female spots in your genealogy. Males are doing more double duty than females.

As in one man getting multiple women pregnant?


Yes.

The estimate I'm familiar with also says that historically about 80% of women, compared to 40% of men, successfully reproduce.


That 80% figure feels very high. I suppose it depends on how a women is counted first of all(for example capable of reproduction). But then how many of them actually died in child birthing process without producing any off-spring that survived.


It's more or less correct because you must remember people don't live forever. Every time you go back more than a couple generations you're starting back at 0.

It's why evolution is such a no brainer. Think of a creature with just a few years of life. They have countless ancestors every century, every millenia. Countless millions and billions of parents came before us.


I think if the author read Time Enough for Love, they would understand where their math went wrong.


There was a paper in early 2000s with some mathematical modeling about that, see https://www.scientificamerican.com/article/humans-are-all-mo...

TLDR all Europeans may have a common ancestor as late as 1400 BCE, maybe 55 CE.


I wish they'd teach the charlemagne paradoxon in schools.


Ah, of course it has a name. Thanks for the lead:

* https://www.scientificamerican.com/article/humans-are-all-mo...

* https://www.nationalgeographic.com/science/article/charlemag...

And no they did not teach me that in school ;D



Most people cant name all 8 of their great grand parents.


Such titles, having questions in parentheses somewhere, look unprofessional to me. I don’t know why.


My theory is that your mum is your sister and your grandma?


When I had 23 & Me, it told me that I am 95% Ashkenazi Jewish. I already knew this. What I didn't know is that I am 5% Gentile. This made me very happy, and I wish the percentage were much higher. Some people in the more conservative parts of the Jewish community are dismayed by intermarriage. I say, please more intermarriage. A lot more. Two thousand years of endogamy didn't do us any favors at all.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: