People and North America

[Politesse]

Much better to eat the animals themselves, though that will require killing them.

No small feat, where mammoths are concerned. But our ancestors most certainly managed the feat.

 

[bilby]

Much better to eat the animals themselves, though that will require killing them.

No small feat, where mammoths are concerned. But our ancestors most certainly managed the feat.

...and once you've got the feat, they'll fall over and you can reach the rest of the animal.

 

[Jokodo]

AFAIK the only Y-haplogroup known to be implied for pre-Columbian South America is Q-M1107. The mutation rate of the Y-chromosome is rather well calibrated by now; and it was about 16,000 years ago that Q-M1107 diverged from a sibling clade (Q-L330) found in present-day males from Hungary and Central Asia. (Most of Q-M1107 is in the Q-M3 subclade dated to 14,000 years ago. The second largest subclade is Q-Z780, found among Creek tribesmen and in the Anzick-1 skeleton.)

This does not preclude earlier settlements, of course, but implies that the early settlers -- or at least their agnatic line -- went extinct.

Strictly speaking and without additional information, it only implies that the agnatic lines of anyone who isn't agnatically related to some present day Hungarians and Central Asians went extinct - the ones who are, ie Q-M1107, could easily be the descendent of just those early settlers.

Looking at just two clades and their separation date, Q-L330 could just a easily derive from a reverse migration event, from America to Asia (We do have good evidence such migrations occurred several times throughout the Holocene). It is only by rooting the branch containing the pair in a larger tree most of whose branches are exclusively Asian that this becomes increasingly implausible. Even so, while slightly less parsimonous, the hypothesis that Q-L330 came from America is still far from outrageous. What *would* be outrageous is postulating that it migrated from America, as did all of its first degree cousins and all of its second degree cousins, with Q-M1107 the sole survivor in the old land.

tl:dr: the most relevant date might not be the separation date from the (singular) closest Asian relative, but the date at which the otherwise mostly Asian branch it belongs to underwent a major radiation.

Source for holocene "back migrations": https://www.science.org/content/art...enes-traveled-back-siberia-new-genomes-reveal

 

[Swammerdami]

AFAIK the only Y-haplogroup known to be implied for pre-Columbian South America is Q-M1107. The mutation rate of the Y-chromosome is rather well calibrated by now; and it was about 16,000 years ago that Q-M1107 diverged from a sibling clade (Q-L330) found in present-day males from Hungary and Central Asia. (Most of Q-M1107 is in the Q-M3 subclade dated to 14,000 years ago. The second largest subclade is Q-Z780, found among Creek tribesmen and in the Anzick-1 skeleton.)

This does not preclude earlier settlements, of course, but implies that the early settlers -- or at least their agnatic line -- went extinct.

...
Looking at just two clades and their separation date, Q-L330 could just a easily derive from a reverse migration event, from America to Asia (We do have good evidence such migrations occurred several times throughout the Holocene). It is only by rooting the branch containing the pair in a larger tree most of whose branches are exclusively Asian that this becomes increasingly implausible.

Obviously I didn't have room to publish the entire Q-haplogroup tree! The common ancestor (Q-L54) of Q-M1107 and Q-L330 is part of a "bushy" tree with wide distribution in Asia.

Although Q-haplogroup is considered "rarish", there are TWO distinct clades of Q found in Scandinavia (and places colonized from Scandinavia, e.g. Iceland and Shetland Islands). The two Scandinavian clades are each only about 3000 years old, but their most recent common ancestor is Q-L56 (20,000 ybp). The less common of the two clades, Q-L804, is dated to 1300 BC and is a subclade of afore-mentioned Q-M1107! This MIGHT be a back-migration like you mention!
 Haplogroup Q-L804 (Y-DNA)

 

[Jokodo]

I might be strolling off topic here, but I find it interesting to note that it is entirely plausible, in the sense of "more likely than not", that every living European, with the possible exception of recent immigrants from Africa or Oceania and their descendents, has ancestors who lived in North America within the last 2000 years.

My Argument rests on three pillars:

- the evidence for (continuous?) gene flow/ contact across the Bering Strait, and between different parts of Northeast Asia (see the research linked in my last post)
- the identical ancestors point for all Europeans has been modeled to be surprisingly recent, around the year 1000: https://www.nationalgeographic.com/science/article/charlemagnes-dna-and-our-universal-royalty
- Pannonian Avars!

The Avars were a group of steppe nomads who controlled much of East Central Europe (conservatively, Vienna to Odessa) in late antiquity / early middle ages until the joint forces of Bulgars and Franks kicked their asses around the year 800, by which time they had intermingled plenty with the locals and were probably speaking a Slavic dialect. While their precise origins have long been shrouded in mystery, recent genetic research seems to confirm the theory that they are the descendants of the Rouran who controlled East Mongolia and adjacent areas in Manchuria and the Russian Far East until their empire was destroyed in 550 AD (which would make theirs the fastest recorded long distance migration before the 1500s, as they showed up in Hungary by 568): https://www.eva.mpg.de/press/news/article/origins-of-the-avars-elucidated-with-ancient-dna/

Now if all Europeans share the same set of ancestors roughly 1000 years ago, then we can confidently say that all Europeans have plenty of Avars among their ancestors, with a huge error margin.

The weakest pillar is the first one: It's well known that steppe polities readily incorporated individuals of unrelated origin, and while it is certainly plausible that many Avars in 550 had recent ancestors from substantially further East than their heartland. And just a small subset of Avars having a single Alaskan ancestor when they started their westward trek suffices for my argument. However, while a more or less steady trickle from Alaska into Siberia is compatible with the genetic data mentioned earlier (and arguably more plausible from an ethnographic and archeological point of view), so are infequent episodes of more substantial migration with millennia of perfect isolation in between, which could substantially push back the time of Europeans' last Alaskan greatgrandparent.

 

[Swammerdami]

I might be strolling off topic here, but I find it interesting to note that it is entirely plausible, in the sense of "more likely than not", that every living European, with the possible exception of recent immigrants from Africa or Oceania and their descendents, has ancestors who lived in North America within the last 2000 years.

These types of questions are fascinating! For example,

Find the smallest N which satisfies​

Any two living humans on the planet are related as Nth-cousins or closer​

The answer is unknowable, but if we could somehow know the answer I suspect it might be something like."N=56 until a certain old man in the Andaman Islands died 4 years ago. Then it suddenly dropped to N=48."
Google comes up with

the most distant person within your culture or ethnicity is probably closer to you than a 15th cousin, while the farthest relation you have on Earth is likely to be as far as a 50th cousin.

Another question is "What is the typical 'N' (cousin distance) for two typical Americans?"
It would take a good model to come up with a good estimate. Googling for an estimate, I find an assumption that a woman has exactly two children that survive to adulthood. But in practice wasn't the average parameter much higher on the American frontier?

 

[lpetrich]

We may not be able to get any precise number, but from genetics, one can deduce population histories. There has been a lot of work on that, work that gives us upper limits on how many generations of separation.

 Mitochondrial Eve is a woman who lived some 100 - 150 thousand years ago in eastern or southern Africa, and  Y-chromosomal Adam is a man who lived around the same time and the same place.

With a human generation time of roughly 25 - 33.3 years, this is 3,000 - 6,000 generations.

 

[Jokodo]

We may not be able to get any precise number, but from genetics, one can deduce population histories. There has been a lot of work on that, work that gives us upper limits on how many generations of separation.

 Mitochondrial Eve is a woman who lived some 100 - 150 thousand years ago in eastern or southern Africa, and  Y-chromosomal Adam is a man who lived around the same time and the same place.

With a human generation time of roughly 25 - 33.3 years, this is 3,000 - 6,000 generations.

An upper limit is useful when it allows us to exclude values that would otherwise have been plausible. An "upper limit" that is two full orders of magnitude, or somewhere around 8-10 sigma, removed from the most likely value, adds no information whatsoever.

The chance that there was a severe bug in the software used to estimate the age of mitochondrial Eve/Y Adam is orders of magnitude higher than the chance that the real age of the last biparental common ancestor is anywhere near theirs.

 

[Swammerdami]

We may not be able to get any precise number, but from genetics, one can deduce population histories. There has been a lot of work on that, work that gives us upper limits on how many generations of separation.

 Mitochondrial Eve is a woman who lived some 100 - 150 thousand years ago in eastern or southern Africa, and  Y-chromosomal Adam is a man who lived around the same time and the same place.

With a human generation time of roughly 25 - 33.3 years, this is 3,000 - 6,000 generations.

A human has more than a trillion^* 38-great grandparents; most of them lived about 1100 years ago. Of this humongous number, ONLY ONE is the AGNATIC 38-great grandfather -- the ancestor from whom you (or your father) inherited your/his Y-chromosome. And ONLY ONE is the UTERINE 38-great grandmother -- the ancestor from whom you inherited your mitochondrial DNA.

It is VERY easy for conversation to get confused when this distinction isn't treated with care.

(* - It i slightly ridiculous to say you have a trillion ancestors; this is more than the number of humans who have ever lived. What these are are ancestor SLOTS.)

 

[Swammerdami]

The chance that there was a severe bug in the software used to estimate the age of mitochondrial Eve/Y Adam is orders of magnitude higher than the chance that the real age of the last biparental common ancestor is anywhere near theirs.

I don't know about mtDNA, but Y-chromosome mutation rate has been rather well calibrated. DNA from 25,000 year-old skeletons have been sequenced (or partially sequenced) and found to conform to the expected mutation rate. Computed dates for divergence conform to known history or pre-history. The fan-out of the FT haplogroup to various clades all over South Asia, West Asia, SE Asia and Central Asia began 49,000 years ago; in agreement with Wikipedia's "Anatomically modern humans (Homo sapiens) reached Central Asia by 50,000 to 40,000 years ago." Et cetera, et cetera. A large portion of males surnamed Cohen (or its variants) are calibrated to have a common ancestor near 500 BC, about what would be guessed via the Captivity in Babylon. Et cetera.

ETA: The common ancestor of R1b-P312 is dated to 2500 BC, in close agreement with the fan-out of the Bell Beaker culture. BOTH Bell Beaker AND R1b-P312 experienced VERY rapid fan-out. One can almost "read out" the expansion/migration pattern from the Y-haplogroup tree. One can almost locate specific Copper or Bronze Age Kings (although their names are unknowable)!

 

[Jokodo]

The chance that there was a severe bug in the software used to estimate the age of mitochondrial Eve/Y Adam is orders of magnitude higher than the chance that the real age of the last biparental common ancestor is anywhere near theirs.

I don't know about mtDNA, but Y-chromosome mutation rate has been rather well calibrated. DNA from 25,000 year-old skeletons have been sequenced (or partially sequenced) and found to conform to the expected mutation rate. Computed dates for divergence conform to known history or pre-history. The fan-out of the FT haplogroup to various clades all over South Asia, West Asia, SE Asia and Central Asia began 49,000 years ago; in agreement with Wikipedia's "Anatomically modern humans (Homo sapiens) reached Central Asia by 50,000 to 40,000 years ago." Et cetera, et cetera. A large portion of males surnamed Cohen (or its variants) are calibrated to have a common ancestor near 500 BC, about what would be guessed via the Captivity in Babylon. Et cetera.

ETA: The common ancestor of R1b-P312 is dated to 2500 BC, in close agreement with the fan-out of the Bell Beaker culture. BOTH Bell Beaker AND R1b-P312 experienced VERY rapid fan-out. One can almost "read out" the expansion/migration pattern from the Y-haplogroup tree. One can almost locate specific Copper or Bronze Age Kings (although their names are unknowable)!

Yes to all of that. And STILL, the chance that these calculations are way off despite these entirely coincident apparent good fits is orders of magnitude higher than the chance that the last biparental common ancestor lived anywhere near as long ago as y-chromosomal Adam, for pretty much the reasons you pointed out.

 

[Jokodo]

I might be strolling off topic here, but I find it interesting to note that it is entirely plausible, in the sense of "more likely than not", that every living European, with the possible exception of recent immigrants from Africa or Oceania and their descendents, has ancestors who lived in North America within the last 2000 years.

These types of questions are fascinating! For example,

Find the smallest N which satisfies​

Any two living humans on the planet are related as Nth-cousins or closer​

The answer is unknowable,

The answer is "no larger than m such that no living human is more than m+1 generations" removed from the last common ancestor of all humans, and probably substantially, but not orders of magnitude, smaller". That was easy!

" but if we could somehow know the answer I suspect it might be something like."N=56 until a certain old man in the Andaman Islands died 4 years ago. Then it suddenly dropped to N=48.""

I think a lot of the answer depends on whether we believe there is any surviving South Americans without any level of post-Columbian admixture from the old world, orc Polynesian admixture from a few centuries earlier. If there are, I'd bet a limb they are one of the far ends due to mere geography, being on a distant dead end of a network of human inhabited land masses. If there aren't, if everybody has a rapey land owner, a runaway slave, an abducted settler, or a visitor from Rapa Nui among theirs - and I think this is way more likely than not - the overall figure will be substantially lowerv and I'd be much harder pressed to pinpoint the extremes.

I still don't think the Andaman islands would be one of them - they may be isolated, but whenever someone did show up, that someone was from a maritime population operating in the Indian Ocean and thus probably from one of the most cosmopolitan places of the time, immediately knocking down their genealogical distance with practically *everyone* all at once.

 

[Jokodo]

My gut feeling would be that a figure around 50 generations is high under the assumption that every South American who doesn't have Erst African ancestors at least has European ones and vice versa but low of the are still people on the continent to whom the only route its by the Bering Strait.

 

[Swammerdami]

These types of questions are fascinating! For example,

Find the smallest N which satisfies​

Any two living humans on the planet are related as Nth-cousins or closer​

The answer is unknowable,
" but if we could somehow know the answer I suspect it might be something like."N=56 until a certain old man in the Andaman Islands died 4 years ago. Then it suddenly dropped to N=48.""

I think a lot of the answer depends on whether we believe there is any surviving South Americans without any level of post-Columbian admixture from the old world, orc Polynesian admixture from a few centuries earlier. If there are, I'd bet a limb they are one of the far ends due to mere geography, being on a distant dead end of a network of human inhabited land masses.

This is what I thought, when the same topic came up several years ago at another message-board. But the Board's expert on South America insisted that tribes isolated in the depths of the Amazon forest aren't really as isolated as they seem. He insisted -- though without any mathematical model -- that Conquistador blood must have penetrated throughout the continent.

If there aren't, if everybody has a rapey land owner, a runaway slave, an abducted settler, or a visitor from Rapa Nui among theirs - and I think this is way more likely than not - the overall figure will be substantially lowerv and I'd be much harder pressed to pinpoint the extremes.

I still don't think the Andaman islands would be one of them - they may be isolated, but whenever someone did show up, that someone was from a maritime population operating in the Indian Ocean and thus probably from one of the most cosmopolitan places of the time, immediately knocking down their genealogical distance with practically *everyone* all at once.

I think Papua New Guinea may also be a candidate for extreme genetic isolation.

-- -- -- -- -- --

As you go back in time further and further, you reach a point where almost all humans alive then have left us ZERO living descendants. Although the number of pedigree SLOTS doubles with each generation, you reach a point where the number of DISTINCT ancestors starts decreasing. Supposedly this happens MUCH sooner than you might guess:  Pedigree collapse

If one considers as a function of time t the number of a given individual's ancestors who were alive at time t, it is likely that for most individuals this function has a maximum at around 1200 AD.

IIRC, this was (almost?) confirmed explicitly for William the present Prince of Wales, much of whose pedigree is known, going back many centuries.

 

[Loren Pechtel]

One of my history profs was fond of suggesting that the first people to cross into north america were following large herds of animals and subsisting as dung eaters. I can still see him lifting his hand toward his mouth when he made the suggestion, kinda like he was holding a burger. I never encountered the idea elsewhere but he was sure fond of the suggestion. The problem is I never heard of any human population subsisting on dung. We've burned the stuff, but have we ever eaten the stuff?

I have heard of people eating the undigested bits that came through an elephant.

 

[Jokodo]

One of my history profs was fond of suggesting that the first people to cross into north america were following large herds of animals and subsisting as dung eaters. I can still see him lifting his hand toward his mouth when he made the suggestion, kinda like he was holding a burger. I never encountered the idea elsewhere but he was sure fond of the suggestion. The problem is I never heard of any human population subsisting on dung. We've burned the stuff, but have we ever eaten the stuff?

I have heard of people eating the undigested bits that came through an elephant.

And mammoths are just furry elephants, right?

As a famine survival strategy or a a regular foraging strategy, though?

 

[Jokodo]

These types of questions are fascinating! For example,

Find the smallest N which satisfies​

Any two living humans on the planet are related as Nth-cousins or closer​

The answer is unknowable,
" but if we could somehow know the answer I suspect it might be something like."N=56 until a certain old man in the Andaman Islands died 4 years ago. Then it suddenly dropped to N=48.""

I think a lot of the answer depends on whether we believe there is any surviving South Americans without any level of post-Columbian admixture from the old world, orc Polynesian admixture from a few centuries earlier. If there are, I'd bet a limb they are one of the far ends due to mere geography, being on a distant dead end of a network of human inhabited land masses.

This is what I thought, when the same topic came up several years ago at another message-board. But the Board's expert on South America insisted that tribes isolated in the depths of the Amazon forest aren't really as isolated as they seem. He insisted -- though without any mathematical model -- that Conquistador blood must have penetrated throughout the continent.

If there aren't, if everybody has a rapey land owner, a runaway slave, an abducted settler, or a visitor from Rapa Nui among theirs - and I think this is way more likely than not - the overall figure will be substantially lowerv and I'd be much harder pressed to pinpoint the extremes.

I still don't think the Andaman islands would be one of them - they may be isolated, but whenever someone did show up, that someone was from a maritime population operating in the Indian Ocean and thus probably from one of the most cosmopolitan places of the time, immediately knocking down their genealogical distance with practically *everyone* all at once.

I think Papua New Guinea may also be a candidate for extreme genetic isolation.

I wouldn't bet on that. The main staple crop in much of the New Guinea highlands is - the sweet potato, native to tropical South America and Mesoamerica. While there has been some debate whether it was brought by Europeans (or by Malays and Moluccans who had gotten it from Europeans) in the 16th or via Polynesia sometime after the 13th century (we know that Polynesians had sweet potatoes by around 1200 and carried them as far out as Hawaii and New Zealand, we just don't know how far West it spread before Europeans), either way its presence in the most remote parts of the island demonstrates that they haven't been as isolated over the last 1000 years as we like to imagine.

What's worse: Apparently the sweet potato wasn't just a drop in replacement for taro, as it has very different demands on its habitat. So while the higher productivity allowed a substantial expansion of the overall population, fields suitable for taro but not sweet potato were given up and many villages deserted. The transformation was so profound that archeologists working on New Guinea have dubbed it the "Ipomoean revolution" (from the Latin word for sweet potato), and indeed most seem to date it to 1600s.

Even if all of that didn't leave linguistic traces as profound as those of the Indo-European and Bantu expansions - what are the chances of all of that happening without significant population movements and mixing?

And that's before we mention that most groups on the island of New Guinea are mandatorily exogamic. That is, at the clan level, they may still be preferentially endogamic at the level of the tribe or linguistic group, but even so, this means any admixture, once it enters a group, spreads faster within it than it otherwise might.

 

[Jokodo]

As you go back in time further and further, you reach a point where almost all humans alive then have left us ZERO living descendants.

I don't think that's true, or even possible, for genealogical descent, except briefly before a catastrophic event where all but one or a few subpopulations were eradicated without any descendants.

Here's why: given the concept of an
 Identical ancestors point, the mirror image, what we might dub the "Identical descendants point" is also implied: a point in the future where everyone alive today who will still have living descendants is the ancestor of everyone alive then. After that point, your genealogical heritage has become immortal, barring the extinction of the human race. Even if only a single breeding pair remains, their descendants will be yours and those of everyone alive today whose line hasn't died out before. The same is true in the past: mostly everyone alive 40k BP who successfully bred and left any descendants whatsoever over the first few generations after their lifetime is likely the ancestor of mostly everybody 30k years ago and of literally everybody today.

In the same vein, even if the Toba eruption did kill off 98% of humanity 70,000 years ago, while this means that most people alive in 70,500 BP or 71,000 BP didn't leave living descendants, the effect is short lived. By, say, 85,000 BP (or later if the survivors were scattered over a larger area), mostly everyone who successfully bred and still had ancestors 71ky BP probably had somebody among the 2% who survived and is thus a universal ancestor of everyone alive today.

Although the number of pedigree SLOTS doubles with each generation, you reach a point where the number of DISTINCT ancestors starts decreasing.

That's certainly true eventually, if only because of you go back far enough in time, your pedigree already includes a majority of the people alive at the time (or rather, of those among them leaving *any* descendants), so the effect of the smaller total population starts to outweigh the effect of a better penetration of that population. For example, if 60% of the global breeding population were your ancestors 1500 years ago, 85% 5000 years ago, and 100% at the beginning of the Holocene, the absolute number of your ancestors may still be largest 1500 years ago.

Supposedly this happens MUCH sooner than you might guess:  Pedigree collapse

If one considers as a function of time t the number of a given individual's ancestors who were alive at time t, it is likely that for most individuals this function has a maximum at around 1200 AD.

Is that supposed to be globally or only looking at Europeans for example? It sounds plausible for the latter - for any living European, it is possibly true that the number of *European* ancestors drops off again before 1200 because their pedigree has reached such a good penetration of the continent's population that population growth becomes the dominant driver. However I suspect they would be adding a *lot* of ancestors on other continents with every generation further in the past from the few isolated transcontinental ancestors they had at the time.

 

[Jokodo]

To add,

a point where almost all humans alive then have left us ZERO living descendants

is certainly a reality when we talk about *genetic* descent - there might even be a point where we all have just *one* genetic ancestor. Just like there is a y- chromosomal Adam and a mitochondrial Eve, there is an Adam or Eve for every gene, every sequence of the human genome: an individual in which all versions of that sequence currently in circulation converge such that every variant, every allele in the human gene pool has been derived from a version he or she had via subsequent mutations.

Sometime before the oldest of such single- gene Adams and Eves (barring vertical gene transfer), there was whole genome Adam, an individual who had ancestral versions of every gene's Last Common Ancestor. I hesitate to speak of a "person" as I suspect this individual might well be older than the last common ancestor with at least some of the great apes. With horizontal gene transfer, it gets even trickier: if via a viral vector, a variant from the ancestors of camels has entered the human gene pool (whether it has reached fixation or not by now) sometime after who would otherwise be our genome wide Adam, that function will become relegated to someone before the last common ancestors of humans and camels...