search this blog

Thursday, December 23, 2021

When did Celtic languages arrive in Britain?

A new paper at Nature by Patterson et al. argues that Celtic languages spread into Britain during the Bronze Age rather than the Iron Age [LINK]. This argument is based on the observation that there was a large-scale shift in deep ancestry proportions in Britain during the Bronze Age.

In particular, the ratio of Early European Farmer (EEF) ancestry increased significantly in what is now England during the Late Bronze Age (LBA). On the other hand, the English Iron Age was a much more stable period in this context.

I don't have any strong opinions about the spread of Celtic languages into Britain, and Patterson et al. might well be correct, but their argument is potentially flawed because:

- significant population shifts need not result in any noticeable changes in ancient ancestry proportions

- ancient ancestry proportions can shift without significant migrations from afar due to cryptic population substructures

- large-scale population shifts need not result in langage shifts, especially if they're gradual

- small-scale population shifts can result in language shifts, especially if they're sudden.

Indeed, when I plot some of the key ancient samples from the paper in my ultra fine scale Principal Component Analyses (PCA) of Northern and Western Europe, it appears that it's only the Early Iron Age (EIA) population from England that overlaps significantly with a roughly contemporaneous group from nearby Celtic-speaking continental Europe. The relevant PCA data are available here and here, respectively.

See also...

Celtic vs Germanic Europe

Avalon vs Valhalla revisited

R1a vs R1b in third millennium BCE central Europe


1 – 200 of 434   Newer›   Newest»
vAsiSTha said...

793 samples from this paper added to map with metadata.

Steppe said...

I have a question on a different topic, unfortunately I can no longer get into Anthrogenica although I have reset my password and wrote to administrators, no message returned, nothing, someone might have the opportunity to contact Anthrogenica, I am known to Thread as Alain and thanks again to Davidski that everyone has the chance to read the latest studies on Eurogenesblog!

vAsiSTha said...

If Celtic arrived post 1000bce then what was the language spoken by the 1500bce bell beaker like people?

Davidski said...

Something that went extinct sooner or later.

Wouldn't be the first time.

Carlos Aramayo said...

Regarding Patterson et al. paper, take a look at the article in BBC:

"The mass-movement of people originated in continental Europe and occurred between 1,400 BC and 870 BC. The discovery helps to explain the genetic make-up of many present-day people in Britain. Around half the ancestry of later populations in England and Wales comes from these migrants. It's unclear what caused the influx of people during the Middle to Late Bronze Age, but the migrants introduced new ritual practices to Britain. The results, published in the peer-reviewed journal Nature, are based on DNA extracted from 793 ancient skeletons..."

vAsiSTha said...

"Between 1000 and 875 BC, EEF ancestry increased in southern Britain (England and Wales) but not northern Britain (Scotland) due to incorporation of migrants who arrived at this time and over previous centuries, and who were genetically most similar to ancient individuals from France."

I'm confused. Don't we already see bellbeakers in British Isles post 2000bce? Are these new people different?

Rafs said...

vAsiSTha: Perhaps alteuropäisch? Alteuropäisch seems like a good candidate for the language that the Bell Beakers spoke.

Bob said...

Great observation David.
The Patterson uses high EEF individuals at Cliffs End and East Kent Access Road (adjacent and next to Ramsgate - where the ferry leaves for the short crossing to France) to show Knoviz Urnfield like ancestry to model the Late Bronze Age migration into southern England.
There are a number of background issues to cover here.
In the LBA that area would have been a small island and landing/trading post - servicing the whole Atlantic Bronze Age network. Isotope data - and plotting on the NW Europe G25 - shows a large variety of non-local people.
While the highest EEF individuals may represent Urnfield like ancestry, they may actually be travellers / traders and not typical members of the source population that migrated into England in the LBA. (While this is not necessarily a problem for use in the modelling, it may result in an unhelpful mindset in considering the results.)

Bob said...

So, what is the LBA context?
In simplified terms, to the North of the Rhine were the descendants of the Tumulus culture. Lets put them to one side for the moment. The two main players we are concerned with are the Atlantic Bronze Age community that sat on both sides of the English Channel to include northern France and into the southern Rhine. Genetically, they had a large R1b>L21 male component - but needs further sampling in France to confirm the detail. Archaeologically they had post Deverel Rimbury plainware pottery and traded with their Urnfield cousins along a trading zone - lets say 100km into France. Their Urnfield cousins, many descendants of Continental Alpine Beaker, had a large proportion of R1b>U152 males - as found in samples in Patterson 2021. Autosomally, they have a significant EEF shift and the Knoviz Urnfield samples sit significantly below the Grand Est IA2 samples on the NW Europe PCA.
In the Late Bronze Age, the Urnfield group expanded towards the channel coast - displacing(?) the Atlantic Bronze Age population from this area. So, they either killed their L21 cousins, or merged with them, or displaced them to the coast and back into the Isles.

Bob said...

You may recall presentations, more than a year ago, suggesting LBA migration into southern England may have been just females because all the Y DNA was still R1b>L21 "British". This was based on the false assumption that L21 was formed in the Isles and only existed in the Isles. The L21 TMRCA makes it almost certain that it formed on the continent - perhaps in the Rhine Valley (??)
There is no need for a female migration in to the Isles as the Atlantic population that migrated would have likely been dominated by L21 males. Indeed, isotope data showing migrants, and increased EEF also occurs in L21 males.

So, a simple explanation for your limited EEF LBA shift is that it was the continental element of the Atlantic BA that went back into the Isles spreading the plainware pottery style across southern England.

But what about your EIA shift to greater EEF (when no significant migration was supposed to be happening - according to Patterson 2021)..........

Dumpling said...

@vasistha i dont know. Dravidian?

Bob said...

The EIA sampling is problematic. There are very few EIA samples available in England, and many of them are just bone fragments from Midden deposits. So, we must consider the possibility of a sampling issue.

Putting that to one side.... there are major changes that occur in England from 900 BCE and in particular from 800-650 BCE. Needham (2017) wrote a paper called 800 BCE The Great Divide. In simple terms, new pottery arrived - decorated ware - with features very similar to Urnfield pottery. This was combined with new types of metal deposition (Llyn Fawr) and new types of settlements with Midden deposits. These settlements continued in Wessex to south Wales, right through to the Middle Bronze Age.

Could this be related to the arrival of further migrants with even higher EEF levels? If truely from the Urnfield group, we may expect to see R1b>U152 males more commonly and samples with very high EEF? Well, we have one exceptionally high EEF sample from Rowbarrow on the River Avon in Wessex. This sample sits well within the Knoviz samples on the NW Europe PCA ('below' the Grand Est IA 2 samples). But we have no evidence of a surge in U152 males.

So, the jury is out and we need to look more closely?

Bob said...

What about the Middle Iron Age / Grand Est 2 / La Tene.
Well, La Tene formed and expanded around the Middle Rhine / Grand Est and into the Marne, Oise towards Paris.
In England, there are many MIA sites with La Tene 'characteristics'. Movement of ideas and / or movement of people? Interestingly, there is a greater diversity of Y lineages in the English MIA samples. This included U152 and other potential 'exotic' lineages.

I have not yet systematically looked at the MIA sites in relation to Grand Est 2 - but your PCA does indicate they are reduced EEF compared to EIA English samples (but limited samples for EIA).

So, may close scrutiny of the new samples we have, tell us anything about the English MIA populations and the possibility of migration??
The issue is that the source population would not be a simple Urnfield or Atlantic BA population because a potential La Tene source population is likely a complex mix that we need to define in genetic terms (we are seriouslylacking source population data). It may even include descendants of our Tumulus descendants from north of the Rhine (now occupying the Lower Rhine) as we also begin to get R1b>U106 appearing in England.

So, all far too complex for the Patterson paper to take on (when trying to give a simple big picture message). And it was not willing to address such issues without more detailed source population data from France/ Middle/Lower Rhine.


Q. Can careful analysis of the Patterson samples via NW Europe G25 and Celtic-Germanic PCA begin to shine light into this darkness?
Q. Has anybody got the statistical analysis skills to back up any subjective observations - to support the validity of any resulting conclusions?

Over to all of you..... if you are interested?

Copper Axe said...


I'll ask around, cheers mate.

Simon_W said...

Striking the change between Netherlands_EIA and Netherlands_LIA:

Distance to: Netherlands_EIA
0.03460898 NLD_BA
0.03716847 Scotland_MBA
0.03766227 England_MBA
0.03913399 Scotland_LBA
0.04078435 England_EMBA

Distance to: Netherlands_LIA
0.02394821 VK2020_NOR_North_VA
0.02480505 DEU_MA_Alemannic
0.02599421 ISL_Viking_Age_Pre_Christian
0.02701937 England_Saxon
0.02800401 VK2020_DNK_Funen_VA

This was probably paralleled by the intruduction of Germanic speech.

Another random observation:

Distance to: Austria_IA_LaTene
0.01941866 Italian_Trentino-Alto-Adige
0.01990707 Spanish_Barcelones
0.02128280 French_Auvergne

LaTene people from Austria pretty close to modern inhabitants of Trentino-Alto-Adige, but also quite Barcelonian-like. And this long before any Roman influence reached Austria.

Urki said...

This reminds me of the 2019 Olalde et al papel about Iberia. They also detected a second wave of People coming from central Europe and reaching Spain during the first millenium BCE. These May have introduced celtic languages in Spain. And yes, we don't Know the language of the previous 3rd millenium wave of steppe People that was so important in the genetic makeup of Spaniards.

Samuel Andrews said...

The Celts is important for understanding northwest Europe.

Because they are why most northwest Europeans today are more "southern", than Bronze age northwest Europe.

Bronze age Britain and Germanic peoples, were very "northern." Celtic admixture explains the southern shift away from them in modern northwest Europe.

Most modern Germans have significant Celtic, whether it be French or not, admixture. So this is why they are more southern than the Lombard DNA samples.

Samuel Andrews said...

Iron age Britain also shows a handful of G2a2b and R1b U152.

This is Celtic Y DNA imprint. So, yeah, no more about Celtic Y DNA not being visible.

Samuel Andrews said...

Bronze age Southern Germany & North France shows "low levels" of Kurgan ancestry.

40% on average. They have significant extra layer of Neolithic Swiss/French ancestry.

This is the arena proto-Celts must have formed in. They were the same as Bronze age Brits, just they had more local Neolithic ancestry hence less Kurgan ancestry.

Samuel Andrews said...

Celtic Y DNA on continental Europe is overwhelmingly R1b P312 in every country.

This backs up Bell Beaker origins of Celtic languages.

It is overwhelmingly U152. But there's some DF27. As well some U106.

Rob said...

Well they did say, Late Bronze Age. Seems reasonable suggestion.

Samuel Andrews said...


Interesting. So maybe Netherlands into early Iron age was a continuation of Bell Beaker. Then Germans, who were super similar, came.

There's a Bell beaker R1b U106 in Netherlands, a Middle Bronze age R1b U106, and an iron age R1b U106.

We already knew U106 in Netherlands isn't all of Germanic origin.

Davidski said...

It's reasonable, but is it actually true?

There seems to be something interesting happening during the EIA, and potentially relevant to the Celtic expansions.

But it only shows up in fine scale PCA.

Andrzejewski said...

@Samuel Andrews So the AS migrants reintroduced the “northern” ancestry component into England and Northern Scotland?

Ryan said...

This would be the Urnfield culture intruding, no? Makes sense to me. If you look at the extent of their spread it matches well with where Italo-Celtic would likely be at the time. I think they're the first "definitively" IE-speaking culture in much of western Europe IMHO. Whatever the Bell Beakers spoke could be IE, but it could just as easily be Vasconic IMHO.

Ryan said...

I'd add that the archaeologists at the Navan Fort site near Armagh were of the same opinion as to the timing of the arrival of Celts to Ireland as this paper suggests for Britain. (I went there in November and it was great).

Rob said...

@ Davidski

''There seems to be something interesting happening during the EIA, and potentially relevant to the Celtic expansions.'

It seems constant/ steady rise in 'EEF' over time, consistent with ongoing small scale migration from continent, although the pace might have varied.
What happens when you add Knoviz, Celtic La Tene

Other comments

- great lineage diversity in Bohemia, as expected
- Croat IA - full of J2b2, continuing the trend seen in MLBA
- an E-V13 in Vekerzug (Scythian-influence group of northern Carpathian basin)
- common Slavic assoc. I2-L621 still missing, despite plenty of hg I, incl. close 'Irish cousins' L161

Davidski said...

That steady rise in EEF might not even be relevant to Celtic expansions.

The PCA I posted aren't dependent on EEF or any other deep ancestry proportions, but on more recent drift.

Tigran said...

How many extinct branches of IE do we think there are now?

Rob said...

Right. How else can it be evaluated ? IBD ? RARECOAL ?

Davidski said...

Probably IBD and haplotypes in general.

The differences between those more continental-like England_EIA samples and England_LBA samples are very small, but they're definitely from different populations, potentially speaking very different languages.

Andrzejewski said...

@Davidski “ potentially speaking very different languages.”

You mean very different *IE* languages, correct?

Davidski said...

Yes, IE.

Rob said...

there's a I10228 + sample 'Croat MLBA'.
However his ancestry is 60% Avar-Slav Szolad like
No C14 in study. Must be a medieval Croat

Bob said...

@Rob @Davidski
Excuse my total ignorance in these matters but what is IBD and how might it be useful?

Is there a methodology that could be used for comparing the amount of Grand Est 2 like ancestry (also perhaps Knoviz like) across the LBA EIA MIA samples (so able to identify e.g. high 'Grand Est 2' (or high Knoviz) as an alternative to the high EEF in the paper etc).

As an aside, I find EIA sample I19867 from Rowbarrow interesting as it plots in the Knoviz group. Ancient pop av closest distance to Iberia Cen CA Step then Cz Hallstatt Bylany.

Steppe said...

Here again a very good statement on the spread of the Celtic language in Great Britain by Ian Armit
The results help to shed light on a debate about when Celtic languages ​​were first spoken in Great Britain.The most established theory, based on the analysis of ancient object styles, is that Celtic languages ​​emerged during the Iron Age with Celtic speakers from continental Europe."
But the new evidence supports a competing idea, based on linguistic studies, that Celtic languages ​​expanded into Britain earlier, in the mid to late Bronze Age. However, we cannot tell what language a person spoke from a person's DNA, says Armit. "

thus linguistics and genetics speak a different language again, it could be that early Hallstatt groups from the western circle brought the Celtic languages ​​to Great Britain as traders in the late Bronze Age and it would later be expanded by Celtic groups in the Iron Age, as the Celts had their origins predominantly in the western Hallstatt culture.

ulfing said...

Yeah, I noticed that. The sample plots so close to Balkan moderns that I actually suspected it's a contamination.

Wise dragon said...

@Samuel Andrews

"Most modern Germans have significant Celtic, whether it be French or not, admixture. So this is why they are more southern than the Lombard DNA samples."

Although, I agree with what you've said. Let me add this to your comment. Northern Germans are genetically speaking still Scandinavian-like.

DragonHermit said...

Very interesting that Croat sample. Thats during the Illyrian time period. Another BA/IA west Balkans-related I1 sample. Messapian/Daunians also had I1 IIRC.

I assumed all I1 in the Balkans was Gothic in origin, but this certainly predates that. Funnily enough it's actually been more common than EV13 in West Balkan populations so far.

epoch said...

@Samuel Andrews + Simon_W

That is very interesting. What need to be understood is that these samples are not from the north. We have no samples and hardly any sign of burials in the north.

On account of BB continuity until the iron age: The well know large barrow structure called the Epe-Niersen line (on the Veluwe) was used until the Iron Age.

DragonHermit said...

This paper did more for the Balkans than Britain. Historians have long speculated whether Illyrians were one people or a mixture of unrelated BA/IA cultures that Greeks/Romans grouped together for convenience, used similarly to the term “Scythian”.

No, the Greeks and Romans were correct. This definitively proves the Illyrians were one related group of people stretching from Greece to Austria, marked mainly by J2B2-L283 with some R1B/Z2103 and I1. Very similar to Messapian samples, which historians relate to these people. It’s amazing how well structured the JL283 spread is with Croatian/Slovenian Illyrian samples being closely related uncles/cousins with modern day Albanian samples. The south Dalmatian J/Z638 starts to look more like modern Albanian samples, which geographically makes perfect sense.

It’s time to rethink the J2B2-L283 as an IE marker, especially since it’s associated with Z2103. We have 0 samples of R1bL51 in the steppe, yet it’s considered an IE marker. The same as R1a in the Yamnaya, This is clearly an Indo-European marker, as such a patriarchal IE-speaking people with some steppe heavy samples could not be like G. We have J from the ancient steppes along with R1a and R1B.

epoch said...


There are several theories of IE substrates, e.g. the Northwestblock theory.

Matt said...

If ancestry proportions are truly homogenised between LBA and IA, and the sample set and available SNPs are large then that should make some formal tests easy.

Of the form:

F4(England_EIA, England LBA; French, German)
F4(England_EIA, England LBA; Irish, French)

And so on.

That might not provide evidence of the directionality if any significant stat is found, but if they are non-sig that tells us something.

They should be able to identify some individuals, out of such a large set, for inclusion in their ultra-high coverage set . Then the haplotype chunk sharing or rare variant models might become more doable, with decent sized robust sets. It doesn't seem immediately promising given Cassidy's earlier study which gave chunk sharing with Rathlin BA only really showed a discontinuity there with England, not Wales, and most plausibly due to Angles and Saxons; but something might show up.

Davidski said...


IDB = Identity by Descent. Genealogical links via haplotype sharing.

But even smaller haplotypes that aren't IBD could be useful.

The point I'm making here is that looking for population shifts in a homogeneous gene pool like BA/IA Western Europe might not be possible by comparing deep ancient ancestry proportions, simply because basically everyone will have the same deep ancestry proportions.

So whatever method we choose, we need to look at more recent, fine scale ancestry components or recent genetic drift.

epoch said...


The Dutch early Iron Age samples are from around Nijmegen whereas the late samples are for North-Holland and Zeeland.

Matt said...

Extended Data Figure 6 - - does make it clear how reliant the finding of the pulse is on the samples from Kent...

Rob said...

If these high EEF continentals come from Cliif's End pit and Margetts Pitt, they dont seem to be 'conquering Celts' or 'Urnfelders'. A couple are teenagers, one an old women, some are fragmentary remains.

Andrzejewski said...

@Samuel Andrews & wise dragon “Although, I agree with what you've said. Let me add this to your comment. Northern Germans are genetically speaking still Scandinavian-like.”

Yes, it explains the differences between Northern Germany, historically “Protestant” and dominated by Prussia vis-a-vis Bavaria and Austria, historically “Catholic”. But besides cultural and lifestyle differences it surely helps to explain how come Germans from the Rhine regions like Koblenz are much more swarthy on average than their northern neighbors: it’s that extra Middle Eastern Barcin farmer + darker skinned WHG natives’ contribution to their gene pool (which made them that way).

Tigran said...

But is this I1 likely to be ancestral to modern Balkan I1?

Tom said...

Lack of a properly sampled France is bound to skew things


You are talking about I-Y3120 which is I2, not I1. The oldest I1 in the Balkans are the Germanic samples from the Roman limes paper. This paper did not find any I1 in that area.

Samuel Andrews said...


Yeah. The slight "racial" difference definitely exists between northern & southern Germany.

I also remember both Tacitus and Ptolmey said German tribes were set apart from neighboring nations by higher amounts of blonde/red hair and blue eyes.

Tacitus was a top Roman historian. Ptolmey was a top Roman geographer. They both took time to mention physical appearance when discussing Germans. People were innocent back then and weren't called racist for seeing pigmentation as a defining trait of populations.

Some people say we're racist, it is bull shit.

Tacitus even went as far as to say Germans based on their physical appetence must be a pure, indigenous race untouched by none-European geneflow via the Mediterranean sea. He would be called an anti-immigration white nationalist today. Even though he was making innocent observations.

It makes sense now that we have ancient DNA. Germans were significantly more northern genetically than Celts. So would have been noticeable more blonde.

What it looks like, is Germans were a far northern nation who migrated south so it makes sense they stood out. They weren't indigenous like Tacitus said. They looked out of place because they came from further north.

Wise dragon said...

@Samuel Andrews, @ Andrzejewski

Here's the thing, Northern Germans indeed exhibit a higher frequency of blondism and blue eyes than Southern Germans. Real brunettes among native Northern Germans are also quite rare. With that being said, the swarthiness of Southern Germans is a bit exaggerated since they have a fair amount of blond hair and blue eyes, but many Southern Germans have chocolate brown hair and eyes, too. Nevertheless, these Southern Germans with dark hair and eyes are usually pale and basically never olive-skinned. Anyway, modern Germans are strongly selecting against blond hair and blue eyes. For example, very Nordic-looking Germans, quite often have either no kids or children with non-whites or very swarthy Europeans only. I personally know plenty of German males who look like stereotypical Vikings who have kids with East Asian, Latina, and African women, while many blond women have kids with Africans, Muslims, or as I said extremely dark European men. Plus, you can see these diverse couples on the streets and in cities when you walk around on a regular basis. In present-day Germany, dark hair and skin tone are considered more desirable than blond hair and blue eyes. This is not really surprising since Commercials and TV shows, popular culture promote interracial dating, exotic people very much, thus they play a role in the strong rise of interracial dating in Germany, too. So, the future Germans will be less blond and blue-eyed and swarthy.

Rob said...

Bavarians aren’t “swarthy”. I’ve met plenty of bavarians who are stereotypically Germanic. Because they come from the Elbe (being ancient Lombards, Thuringians, and Baiovarii from Bohemia).
Anyhow , I’m more interested in skin phenotype because of medical implications (skin cancer, vitamin D) than aesthetics

Steppe said...

I come from Germany, I can only say that some southern Germans from the Rhineland or Bavaria have a slightly darker complexion (very light brown) (German descent without any other ethnic admixture from Middle Eastern or African countries) but the term "white" is stretchy, so you can look at white from pink, medium-light skin to slightly Mediterranean brown. But there are also people with dark eyes and dark hair in East Friesland, but the so-called "blondism" points more to the north but of course there are also Bavaria blond people only that the brunette is higher and for example in East Germany you can find many phenotypes of hair and Eyes best example Simone Thomalla (brunette type) and Sophia Thomalla (more Nordic type). Southwest Germany has a lighter Mediterranean influence (remnants from Roman times, French and Spanish settlers in the Middle Ages ...) and East Germany, for example, has a Slavic admixture from the early Middle Ages. But there are also Middle Eastern people who look very white and look pale than a Bavarian, but you can mostly tell from his facial features that he does not come from Europe, but I also knew a lady from Iran who looked very Eastern European (you could perhaps also be Indo-European (See CWC-Sintashta-Andronovo remnants of the early Indo-Iranian peoples)

Romulus said...

Merry Christmas Davidski and everyone else too.

Ric Hern said...

@ Romulus

Merry Christmas to you.

Davidski said...

Enough already about Bavarians and their pigmentation.

Assuwatama said...

Iron Age East Europeans were speaking Iranian languages. These Iron Age East Europeans (pontic Caspien Scythians) were partially to wholly descended from Andronovo (Proto-scythian).

This East European steppe ancestry was later brought into India by Scythians Parathians and Persians.

~490bce Darius inscription:

Bāxtriš Suguda Uvārazmiš
Zraka Harauvatiš Θataguš Gadāra
Hiduš Sakā haumavargā Sakā
tigraxaudā Bābirauš Aϑurā
Arabāya (Arabia) Mudrāya (Egypt) Armina (Armenia)

Scythians (soma drinking and Pointy hats) mentioned next to Hindush (Indus)

Scythians were largely R1a1 I guess.

Aryan speaking Mitannis were from Indus. Their migrations coincide the introduction of elephants and zebu bull in Near East. If Harappans were Dravidian speakers We would have instead found Dravidian attestation in Near East around that time.

Sumerian legends about Aratta also point to an Aryan-Sanskrit presence somewhere in East someplace that had control of lapis Lazuli (shortugai-Afghanistan - harappan site) and the artisans who worked in such precious stones by 2100bce.

Lord of Aratta was named Ensuhkeshdanna/Ensuhgirana (sukhesh/Sughir) both are Indian sounding names.

From the Sanauli which some scholars claim as a site of proto-Indo-Iranian migration show no similarity either in pottery or burials of Andronovo.

We have found peacock and camel motifs in Sanauli. There are 0 peacocks in steppes and Bactrian camel has double hump while Camel motif has single hump.

epoch said...

First of all, Merry Christmas.

The sample I11631 from Tiel-Medel-De Reth is not in the G25, most likely because of its very low coverage. That is a pity though, as it is the only one which we could call proto-Batavian. It's dated 95 calBCE - 109 calCE. There is some discussion amongst archaeologists on the nature of the Bavarians (Germanic, Celtic or Germanics settled among Celts) so it would be nice to see what it did, with the caveat that it's a low coverage sample.

Romulus said...

From this paper:

Molenaarsgraaf, Zuid-Holland, Netherlands
In 1966 and 1967 a small Late Neolithic–Early Bronze Age settlement was excavated (h 1967/1.) on the Schoonrewoerd stream ridge near Molenaarsgraaf, in the Rhine-Meuse Delta. Apart from one or two possible house plans, three human burials and an ox-burial were found. These burials have become famous in Dutch archaeology because they were well-preserved inhumations and represent clear examples of people buried ‘Bell Beaker style’ in flat graves in a settlement. The layers around the house contained many Bell Beaker and Barbed Wire Beaker potsherds, indicating occupation dating to 2200–1900 BCE.

Petrous bones from three individuals were successfully analysed for aDNA. The first was from an individual aged c. 15 years (Skeleton 1), who was laid on his left side in a crouched position, facing south, in Grave I (dug right next to house I), and yielded sample I13025 (male) R1b-U106+ , dating to 3635±40 BP (GrN-5131; 2136–1892 cal BCE; Louwe Kooijmans 1974). This accords well with a late Veluvian style Bell Beaker positioned at his feet. A fin-ray found in/near his throat may have been the cause of his death. The second was from Grave II, which has become famous because it appears to be the grave of a ‘fisherman’, containing three bone fishhooks, some flint tools and an antler tool, possibly used to lift fish-traps (Louwe Kooijmans 1974). The 18–24 year old individual (Skeleton 2) in this grave lay in a crouched position on the left side, facing west, and yielded sample I13026 (male) R1b-Z2103+ , dating to 3630±40 BP (GrN-5566; 2135–1890 cal BCE). Recent research indicates that both individuals, but especially Skeleton 2 (sample I13026), probably suffered from vitamin D deficiency during childhood (Veselka 2018).

The third individual (Skeleton 3) probably was only 1.5 years old and yielded sample I13027 (female). She was placed on her right side in a pit that was far larger than the body. A small nail-decorated atypical Beaker was added as a grave good (Louwe Kooijmans 1974, 261). One of the legs showed signs of deformation, probably due to vitamin D deficiency (Veselka 2018). The body is dated to 3700±25 BP (PSUAMS-7847; 2197–1983 cal BCE), which is in the same range as the other individuals from Molenaarsgraaf. They all seem to represent an occupation phase between c. 2150 and 1980 cal BCE, in line with the Bell Beaker pottery found in the graves.

Arza said...

@ Davidski

Apparently there are 77 unintentionally published Unetice samples in the v50 dataset that aren't present in the dataset uploaded to Edmond by Papac. All of them are from the Mikulovice site and almost all of them belong to four families.

Here are their details from the anno file:

Can you add them to G25?

Davidski said...

Can you put them in a Plink file and mail them over?

Arza said...

Sent. They're renamed to Mikulovice_published and Mikulovice_unpublished.

Assuwatama said...

Guys ain't it funny that Vedic deities get attested in near East instead of Andronovo or related Scythian cultures who scholars claim to be the Proto-Indo-Iranian, Indo-Iranian and Iranian urheimat.

Bugaš, Maruttaš, Šuriaš (kassites)
ú-ru-wa-na, mi-it-ra, in-da-ra, na-ša-ti-ya-an-na (Mitanni)

Their Vedic equivalents would be;

Baga Marut Surya
Varuna Mitra Indra Nasatiya

Davidski said...

I'm not aware of any Andronovo texts, so there's nothing to attest.

But those Indo-Iranian phrases sound sort of Balto-Slavic.

And of course the obvious link between Balto-Slavs and Indo-Iranians is Andronovo and Corded Ware.

Bob said...

IDB = Identity by Descent. Genealogical links via haplotype sharing........
.....So whatever method we choose, we need to look at more recent, fine scale ancestry components or recent genetic drift.

Thanks, much clearer for me now.
Seasons greetings to all.

EL said...

@ Dragonhermit That's not right. You can find the haplogroup assignments in the supplementary material. Out of the 793 individuals in this study, there is only 1 sample of I1. It's a Scandinavian-like individual, Iron Age sample I16453. In the supplementary it says that the burial is believed to be Iron Age but that the date is uncertain without supporting materials. There is no ancient I1 from Croatia at all in any study so far, by the way

Bob said...

Re. sample I13026 potentially very exciting sample if Z2103. However, relatively low number of Y reads. Review by FTDNA only confirmed it as M269. Alex Williamson has not so far confirmed a Z2103 call.
Perhaps needs a close look.

Alex and FTDNA have also come up with the U106 call for I13025. (It has much higher Y read numbers)

In general, there are quite a number of samples where Alex Williamson and FTDNA have suggested revisions to the calls published in the paper.

Tigran said...


Doesn't matter. Mittani were from Central Asia and Eastern Europe ultimately. Its time to accept the truth and stop grasping at straws and obscure arguments for why Indo-Iranians weren't from Europe.

Romulus said...

If Beakers were suffering from Vitamin D deficiency, as that excerpt I pasted above stated, then that would be one very obvious factor as to what drove selection for lighter skin during their epoch. Especially if as most of you state that the Dutch Beakers are the population ancestral to all other Beakers. But why in the Beaker era and not before? Neolithic farmers had fairly light skin , they were homozygote for SLC24A5 but had low frequency for SLC45A2 (~25%) compared to Beakers who had it at (~70%)(As per this paper). During the Northern European Neolithic it was much warmer than it is today. They received enough sunshine to grow grapes in Scandinavia. It also looks like it may be the same case for Lactase Persistence.

Davidski said...


This is boring. We've discussed all of these issues before here.

You're making newbie assumptions and mistakes.

CHG Chad said...

Iron age central Europe..Celts or not they seem to be very close to north Italians/Iberians and French overall.The Germanic and the Slavic expansion change the genetics of central europeans and made them plot more northern and norteastern.More samples from Bavaria,Baden,Rhineland,Alscatia and Hesse will help to understand better the genetics of central Europe.But its obvious Celts were not like Germanics or Balto-Slavs genetically.

What happened in LBA/IA Britain is still on debate.

English are more southern shifted from Irish.Why this?Simply because English have continental European admix(from Celts,Normans,French whatever) while Irish are way more Beaker derived.Welsh and Scottish are between them with some individuals being more southern while others plot more northern.What impact the Anglo-Saxons had is also a debate since these people would had been very close genetically with the Beakers.This can be seen only from yDNA lineages and not by autosomal.Thought when it comes to looks and phenotypes modern British are nothing Germanic/Scandinavian looking but they stick to their own phenotypes and looks.

The Celtic mix in Britain it seems to be somewhat 30% and not more.And even South English are more northern shifted genetically from Belgians,South Dutch,Northern French and West Germans.Witch means they have more Beaker and maybe Anglo-Saxon admix.

IA Croats(illyrians???) are mostly J2b2 yDNA and they plot with North Italians/Iberians as well.

Andrzejewski said...

@Romulus “ Neolithic farmers had fairly light skin”

Not necessarily: newer models describe both TRB and GAC alike as (moderately) darker skinned.

Rich S. said...

I apologize up front for not having read all the comments yet, but I wanted to mention that there was a major change in the Isles from the older form of Celtic, Q-Celtic, to the younger form, P-Celtic. Q-Celtic survived in what is now Scotland and Ireland, while P-Celtic took root in southern Britain. The Welsh language, for example, is regarded as a descendant of the ancient Brythonic of southern Britain, and is a P-Celtic language.

Isn't it odd that the older Bronze age genetic pattern persisted in those places where the older form of Celtic speech survived, and the newer genetic pattern asserted itself in the same places where the newer form of Celtic intruded?

Andrzejewski said...

@Foxvillager et al. The LBA/IA may have shifted British pop in a more southerly direction, but the AngloSaxon, Danish and Norman (all Germanic) invasions subsequently recalibrated them towards the original Dutch Beaker EBA gene pool.

Irish cannot be a Beaker language because the Celts who came after Beakers replaced their language. It seems like the Beaker spoke a Northwestblock IE language:

Belgian, Pictish, Lusitanian and Ligurian may all be Beaker/NW Block related.

Davidski said...


This isn't really my area, but in my mind the spread of Celtic has always been an Iron Age thing, simply because the change from Bronze to Iron was such a big change and so it's a plausible link with the massive spread of Celtic.

I think that, if anything, the English EIA samples from this paper have backed up this hypothesis, because they look to me like the most obvious link between the continental and British Celts in terms of recent, fine scale ancestry.

But yeah, it's true that in Scotland and especially Ireland, the people retained a more Bronze Age genetic structure, so this might be directly or indirectly linked to Q Celtic.

Romulus said...


I think using Light/Dark skin is a misleading paradigm for Neolithic Farmers vs Beaker/CWC pops. Neolithic farmers had skin tones similar to modern day Southern Europeans. Do modern Southern Europeans have "Dark" skin? Dark compared to Northern Europeans only. Lighter than any other population. Beakers were intermediate between Southern Europeans and Modern/Iron Age Northern Europeans. Any population homozygote for SLC24A5 will look like Near Eastern populations at a minimum kind of skin tone.

Rob said...

@ Dragon Hermit

''It’s time to rethink the J2B2-L283 as an IE marker, especially since it’s associated with Z2103. We have 0 samples of R1bL51 in the steppe, yet it’s considered an IE marker. The same as R1a in the Yamnaya, This is clearly an Indo-European marker, as such a patriarchal IE-speaking people with some steppe heavy samples could not be like G. We have J from the ancient steppes along with R1a and R1B.''

J2b2 has not yet been found on the steppe either, and those variants of J which have surfaced (mesolithic J1) did not expand during the Bronze Age Yamnaya horizon. I would not be surprised if some isolated Majkop-related J2a appears in the lower Don region

By now, everybody knows that Yamnaya lineages were R1b-Z2103 + minority I2a2a1b.
Also, as has been pointed out by others, there is no I1 in Illyrians yet.

CHG Chad said...


I did not said that Irish is a beaker language.I mention that Irish are the most Beaker devired pop in Isles and in the whole EU ofc.They have affected less from the Celts and the Germanics in comparison with other people in British Isles.Their autosomal and their yDNA is mostly associated with these people from the bronze age.Ofc i am not saying that Celts or later Germanics did not left their impact but in lower numbers compared to English,Scottish,Welsh etc.The Celtic language arrived in Ireland probably via Wales.And the whole Celtization was mostly a cultural phenomenon than anything else.

Now about the Anglo-Saxons and the later Germanic waves we need a specific study about it and how much impact they had.Personally,i really doubt they had an impact to the whole England.Their biggest impact was at the eastern parts of the country...where you can see the Germanic lineages being high in numbers.Danes and Normans did not had a big impact to the locals IMO.If Normans left any impact that it would be in the area of London.They were overall a small elite population that ruled the country with their dynasty.

Btw, the Dutch IA samples are also very northern by genetics and is also a debate if they are Beakers or Germanics.Can easily be Nordwestblock remains i guess.

Rob said...

the p/ q-celtic thing is just one sound change of a host of others.

Linguists often order the following in dis-attachments within Celtic family:

1. Celto-Iberian (shares the least features c.f. other Celtic languages)
This would make sense because Iberia was only episodically in contact with rest of west-central Europe- whether one prefers the 'Atlantic cultural koine' or Urnfield C. So a migrant group moved in at one point, then became disconnected. There are varying opinions about the status of Lusitanian.

2. cis-Alpine Celtic. Physically separated much earlier than the Roman sources state. who are instead depicting migration of Galli, not the earliest Italian Celts.

3. ''There are a fair number of innovations which demonstrate that Transalpine Celtic,
Goidelic and Brittonic are to be grouped under a single node on the Celtic family tree''
Where linguists vary is whether Insular Celtic in turn forms a node within or not

IMO, what is more interesting is how Celtic diffused in comparison to Germanic & Slavic. Very different processes in each case.

Andrzejewski said...

Germanic languages are unique in that they encompass certain traits shared with Balti-Slavic branch and others in common with Celtic or “NW Block”. So it could be that CWC are the progenitors of Balto-Slavic and Indo-Iranian (and maybe also Tocharian, Thracian, Dacian and Cimmerian), all having R1a1 as their ydna uniparental marker, where on the other hand the SGC/BBC were the CWC offshoot which became the NW Block, being it Bell Beaker, Unetice, Hallstatt, La Tene etc.

Germanic could be a Nordic BA fusion of Satem R1a1 Balto-Slavic CWC language with a NW Block one, mediated by R1b lineages.

It’s unlikely that the theories pointing to Germanic languages constituting up to 1/3 of their vocabulary of substrate Farmer and/or WHG languages would ever materialize; almost all vocabulary in all IE is Steppe-derived. The reason that Proto-Germanic has so many singular words is it’s position as an amalgam of 2 distinct and divergent branches of the IE tree.

Rob said...

''It’s unlikely that the theories pointing to Germanic languages constituting up to 1/3 of their vocabulary of substrate Farmer and/or WHG languages would ever materialize; almost all vocabulary in all IE is Steppe-derived''

details may vary, but I see no reason to disavow the anaylses of linguists like Kroonen et al
Nordic Bronze Age has significant MNE admixture, Y-hg I1,. Iron Age Germanic groups picked up additional GAC-related I2a2. So the broad picture fits

Davidski said...


Enough already about the PIE homeland.

It was on the Pontic-Caspian steppe. Get used to it.

Andrzejewski said...

@Rob from the Wikipedia entry:

“ More recent treatments of Proto-Germanic tend to reject or simply omit discussion of the Germanic substrate hypothesis. Joseph B. Voyles' Early Germanic Grammar makes no mention of the hypothesis, like many recent publications on the Germanic language family.
Nonetheless, the hypothesis remains popular in some circles, such as the Leiden school of historical linguistics.”

Yep, many etymologies turn out to be IE.

And btw, Slavs are 1/3 GAC but Balto-Slavic languages are almost 100% WSH’s vocabulary.

Assuwatama said...

I have my reservations

We don't have data from India.
Neither the linguistic data nor the genetic data. I would rather wait for more data and decipherement of Harappan script than put things to stone as gospel truths that too in the midst of many unknowns.

12 Indus Periphery samples for a region which likely had 5 million + inhabitants isnt a good sample size.

Indus valley didn't see population replacements of the scale that Europe saw. Harappan ancestry still dominates steppe ancestry 3:1.

Andrzejewski said...

One flaw in Kroonen’s argument points to his thesis that Neolithic languages were homogeneous across their range, and in particular- that GAC and TRB were related to the Tyrrhenian family. We know it was not the case.

More odious is the claim he made about Pelasgian language being a non-IE Barcin-derived one; we simply don’t know what language the so-called “Pelasgian” spoke, or even if they spoke 1 language.

Rob said...

@ Andrze

Firstly, i agree that Neolithic langauges were not monolithic, quite the contrary they must have been diverse.

However, I don't think the putative heavy presence of agricultural terms in Germanic is the same as Germanic substrate hypothesis which deems ''The Germanic substrate hypothesis attempts to explain the purportedly distinctive nature of the Germanic languages within the context of the Indo-European languages. Based on the elements of Common Germanic vocabulary and syntax which do not seem to have cognates in other Indo-European languages, it claims that Proto-Germanic may have been either a creole or a contact language that subsumed a non-Indo-European substrate language''

This is a very different thing to what has been recently proposed. Now, I'm not going to list out words because I'm not an etymologist. But population histories can unlock the underlying dynamics

''And btw, Slavs are 1/3 GAC but Balto-Slavic languages are almost 100% WSH’s vocabulary.''

What do stats base that on ? maybe some indirect admixture, but the Neolithic ancestry in slavic is bound to be more complex than "1/3 GAC". But a key point is the unanimity of R1a-M417 lineages in pre-Balto-Slavic, as it arguably spread of forest-dwelling fisher-hunter-gatherers which did not made a large impact, and those which demographically survived, would have switched en mass to pBS to survive with little substratic impedence.

The situation in north/central Europe is quite different, where we see a significant survival of pre-Corded Ware middle Neolithics, whether in Unetice or NBA. We see that these are mixed societies. A BB supserstrate on top of Battle Axe alone does not suffice IMO

Samuel Andrews said...

Something interesting, is continental Celts had an almost identical Y DNA genepool to Etruscans.

Vast majority R1b U152/P312, with some G2a2.

It is good demonstration that early Romans, Etruscans were normal Western European people not a "Mediterranean" people. They were more related to Celts than to Greeks.

Matt said...

Not really informative of anything new but just as an example of what can be seen in these very large sets:

1) Using Vahaduo, lowest 10 distance between population averages for all available British samples:

Against all pop averages:
Against modern people:

Uses some customized labels which aggregate or split some things differently. In date order.

Can see the change from being closest to other ancients to getting low distances to present day people in British + Irish Isles. The big sample set that I've aggregated under GBR_England_IA (ignoring the distinctions between EIA, MIA, LIA in the paper), has a particularly low distance, as the relatively high quality samples and aggregating the set reduces noise, plus more recent samples should match recent people better (generally).

(Same thing for Czech Republic set:, and Hungary:
2) Also, using only the very highest quality samples (made it through to G25 on genotypes, SNPs>300,000), all samples distance from present-day Welsh:

(Colour scheme is based on y-dna as per the previous time against EEF plots).

We can see the general trend of falling distance over time, with some bumps up during periods in which outliers with a higher distance are coming in (and at the end point, a modest bump up due to the inclusion of the Viking England/Wales samples who have differentiation which is somewhat unrelated to EEF/WHG/Steppe proportions).

In principle this might be doable with a f4-stat like f4(modern Welsh, Ancient ; modern Welsh, Outgroup) to get a formal stat version. If you did a stat like f4(modern Welsh, Ancient ; modern Welsh, modern Polish) as two groups with broadly similar EEF/WHG/Steppe proportion, then you might be able to isolate a trend of samples sharing more *specific* drift with specific present day people over time that was unrelated to EEF/WHG/Steppe proportion.

Ric Hern said...

@ Andrzejewski

You speak as if Bell Beaker people only existed in England. Just a reminder that Bell Beaker people were found as far East as Hungary and Poland. We know that they were found in the Unetice Culture and from areas where the Tumulus Culture originated, basically from the Elbe all the way to Spain and Ireland. So who can tell if they did not influence the Languages which originated in that massive area ? So if not directly from Bell Beaker Culture it could be from Cultures which grew out of them or people related to them....

CHG Chad said...


They are central european people and not western europeans.Northern French,Belgians,South Dutch,West Germans,South English are west europeans.

The Italian-Latin and Roman genetics are a southern version of the central European genetics.High % of EEF but with decent amounts of Steppe and WHG.Unlike Greeks and Sardinians indeed who were more Meditteranid.But even some Celtic people were very close to Iberians and North Italians with high EEF admix.I am waiting samples from Bavaria,Baden,Hesse,Alsace to see how Iron Age Celts were genetically.If they were like the Austrian samples it means that Germanics had indeed a huge impact there since Germans from these provinces/areas are more northern shifted from North Italians,Iberians and to some French groups.

vAsiSTha said...

""Something interesting, is continental Celts had an almost identical Y DNA genepool to Etruscans.

Vast majority R1b U152/P312, with some G2a2.

It is good demonstration that early Romans, Etruscans were normal Western European people not a "Mediterranean" people. They were more related to Celts than to Greeks.""

Therefore, we shouldn't correlate language spoken with uniparentals and make it a hill to die on. Linguistic processes are much more complex than that.

andrew said...

"ancient ancestry proportions can shift without significant migrations from afar due to cryptic population substructures"

We know this didn't happen in Britain. Its ancient ancestry proportions didn't shift except at times of archaeologically and historically documented mass migrations.

"large-scale population shifts need not result in langage shifts, especially if they're gradual"

We know that the Bronze Age population shift in Britain wasn't gradual. And, it is almost certain that there was a language shift at that point. Presumably, the language replaced was part of a language family associated with the mostly Anatolian Neolithic LBK migration of first farmers. But while the Bell Beaker population that replaced it could have been Celtic (although the linguistic time depth and archaeological relics don't really support that hypothesis), it could have been something else and we have no attestation of pre-Celtic language in Britain due to a lack of literacy in Bronze Age Britain and due to a lack of contact between Bronze Age Britain and any literate visitors who could have remarked upon it.

"significant population shifts need not result in any noticeable changes in ancient ancestry proportions"

This could have happened in a Bell Beaker to Celtic shift which is what your data on Iron Age links seems to suggest.

"small-scale population shifts can result in language shifts, especially if they're sudden."

There are certainly precedents for this in British history, such as the Norman invasion which led to the shift from Old English to Middle English. It could have happened in a Bell Beaker to Celtic transition as well.

Rich S. said...

I understand that someone objected that the difference between Q-Celtic and P-Celtic is a single sound change, the implication being that the difference between the two branches of ancient Celtic is trivial. However, surely that single sound change is merely what was used to NAME the two branches and was not meant to be a complete summary of all the differences between them.

As I understand it, the shift from Q-Celtic to P-Celtic is regarded by Celticists as significant and not at all trivial.

Davidski said...


I actually think that the rise in EEF in Britain during the Bronze Age was a leveling out of cryptic substructures from the Copper Age, as well as some sporadic migrations from continental Europe.

Wee e said...

“Between 1000 and 875 BC, EEF ancestry increased in southern Britain (England and Wales) but not northern Britain (Scotland) due
to incorporation of migrants who arrived at this time and over previous centuries, and who were genetically most similar to ancient
individuals from France. These migrants contributed about half the ancestry of Iron Age people of England and Wales….”

Wow. I would have thought there would be more of a stir on a genetics forum about the sheer size of this, genetically speaking.

Also, what do the authors mean by “at this time and over previous centuries”? (Answers in layman’s terms…?.) is the date range because of the availability of samoles limited to that time range?

Also the difference from Scotland. (Implies that maybe the Romans put the stone wall in a zone where existing political/cultural differences were already apparent?)

Some people have wondered whether some Germanic speakers were not in fact already in parts of England well before the Romans left: not counting troops brought by the Romans, This may be sounds a really stupid question: but in France, at 1k bce, what is the evidence for the existence there of “Celtic” language rather than, say, a Germanic or non-indoEuropean one (like more widespread basque). Answers pitched for a layman would be appreciated.

Wee e said...

“As I understand it, the shift from Q-Celtic to P-Celtic is regarded by Celticists as significant and not at all trivial.”

I don’t understand it. It’s one sound shift. There are other differences, of other kinds, that dwarf it. And it seems to matter when people want it to matter, (that Pictish is P Celtic) but not when they don’t (that Pictish stones read Maqq for “son”/of).

The same sound shift even applies within English in the UK. The name of the artist Vincent Van Gogh is pronounced with an “f” in England, for example. (Which amuses and baffles Scots). There are other examples: peek/keek and so on. And other isoglosses in UK English (like chib/shiv): yet these are not seen to be some major, fundamental split.

What makes this PQ significant compared with the satem/centum shift?

P/Q runs across europe, sometimes between languages, sometimes within dialects of what are held to be one language? I can’t find anyone who will explain it to me, why this that makes Maqq (Pictish) of Mab/Map (welsh) at the same time shows that Pictish is P celtic but also enacts this hugely significant between P and Q.

Wee e said...

“the implication being that the difference between the two branches of ancient Celtic is trivial.”

No: the implication being that THIS is a trivial difference. Not at all one that gets to the heart of the differences. Not one that is at all informative, either.

There are some big differences but also some big similarities, * that cut right across PQ.*. Just not in a way that neatly puts Welsh in one category and Irish and Gaelic in another.

The earliest ogham known in Ireland and occasional examples in west Britain, has very little of the broad/slender business and lenition that characterise modern Irish and Gaelic. (The later Pictish Ogham also lacks it).

Syntax and grammar in Welsh and Gaelic are more alike and more conservative than for Irish. Welsh basic everyday vocabulary is about 50% Romanised (ie the Latin rather than Celtic forms for fish and bread etc) : what is there to say that the second-language speakers, or a response to them, did not also bring about the P form?

What is it that says it existed beforehand?

And I am also confused by the insistence that Pictish ogham proves P celtic was the language spoken in north west Scotland, when one of the most common words it contains is “maqq”. A Q that somehow proves it’s a P-language.

Too much of “insular Celtic” linguistics has been founded on circular arguments that ping-ping between archaeology and 19th century linguistics. And completely ignores a wealth of (sigh) Q celtic placenames and name-elements in large swathes of England.

Once again: not that there is a trivial difference: but that PQ is strangely unproductive in working out the relationships between insular languages and their development.

Wee e said...

@Andrew “ We know this didn't happen in Britain. Its ancient ancestry proportions didn't shift except at times of archaeologically and historically documented mass migrations.”

Apparently now we do, though. the paper linked to says that “Between 1000 and 875 BC, EEF ancestry increased in southern Britain (England and Wales) but not northern Britain (Scotland) due to incorporation of migrants who arrived at this time and over previous centuries, and who were genetically most similar to ancient individuals from France. These migrants contributed about half the ancestry of Iron Age people of England and Wales” ….?

I’m not being funny, I just would like to hear more of what people think of the paper itself. (In layman’s terms if possible.)

Davidski has hinted that he thinks this may not be immigration at the time that the authors suppose, but prior (“cryptic”) populations becoming kore widespread or integrated (if I understood him correctly, which I may well not have). Is this what you mean?

And why the difference between north and south?

One thing that comes up my back is knowing that, in Scotland at least, there were quite a lot of very different sorts of funeral types going on for a long time, in a patchwork geographically, some of which normally left no bones by the nature of the rite, and some of which left nine because of the preponderance of very acid soils in much of Scotland. This may hold true in other parts of both islands.

Wee e said...

Doh! Pictish ogham in north EAST Scotland. (Well, just occasionally in the west.)

Rob said...

@ Andrew

''Presumably, the language replaced was part of a language family associated with the mostly Anatolian Neolithic LBK migration of first farmers

Sure this is a hypothetical side-comment, but i would still point out that this is a rather impossible scenario given that LBK never reached Britain. British farmers arrived 1,000 years after LBK ended and were vastly different

@ Vasistha/ Sam

''It is good demonstration that early Romans, Etruscans were normal Western European people not a "Mediterranean" people. They were more related to Celts than to Greeks.""

Therefore, we shouldn't correlate language spoken with uniparentals and make it a hill to die on. Linguistic processes are much more complex than that.''

That’s the final product after establishment of steady state, the correlation is strongest during the expansion phase

I’m not sure eastern Halstatt people from czech R. were ever Celts, anyway. The eastern Halstatt people were deeply intertwined with the Alpine region, north Italy & northwest Balkans; hence we see Etruscan, Raetian and hypothetically other non-IE languages being prominent there. The local variety of IE 'east Alpine IE' probably became extinct

Andrzejewski said...

If Celts had more Sourhern shifted EEF ancestry compared to Proto-Germanic, then how would you explain their language containing almost zero non-IE vocabulary compared to the former?

Samuel Andrews said...


I suppose this is a disagreement on geograhy.

I would consider Germany, Bohemia as apart of Western Europe. This is where the IE ancestors of Italic languages originated.

Likewise I'd consider Celts in Germany & Bohamia as Western European.

Celts in Bohamia were really close to Celts in France. They were a West European people.

Italic tribes were also a West European people.

LivoniaG said...

HI, all! The summary of this new paper on Science Daily has this sentence in it:
"Previous research has shown that large-scale movement often accompanied language changes in pre-state societies."
If you have a copy of the actual paper does it say where this research is from? What "pre-state" societies are they talking about?

Also it seems like cremation as a funeral practice should be traceable as it caught on going back to when? the Neolithic? Is there a map somewhere that shows how cremation or funereal exposure spread through Europe over time. I guess it would be relevant to where aDNA evidence is not available.


KM said...

One thing that is very obvious from the data they have made available is that the Hallstatt and La Tene samples from Czechia/Slovakia/Hungary are remarkably diverse. It's quite striking. They seem to be caught in flagrante mixing with what I presume are (southern) Germanic Bronze Age people, along with sporadic instances of people from further south. The "trade contacts" identified by archaeologists and similarities in art across the swathe of land linking Denmark and Greece might reflect this. I wonder to what extent this was generated by some process of empire and conquest.

George said...


Burial practices in iron age Britain:

Additional references can be found on

CHG Chad said...

@ Samuel

I am talking about genetics here not about geography.

There were Celts like the French and there were Celts being closer to North Italy/Iberia.But the Romans,Latin,Etruscans were not Western Europeans by genetics.Their genetics were southern shifted for being western euro.

Not all of the French are western Europeans.Some of them have central European genetics.With steppe and EEF ratios being balanced.While other French groups are even closer to Iberians.Northern French are indeed western europeans by genetics since they plot closer to Belgians,South Dutch,West Germans and South English.

Italics,Etruscans,Latins(Iron age Italy) had high EEF mix to be considered western europeans.

Dave the Slothtopus said...

@KM "One thing that is very obvious from the data they have made available is that the Hallstatt and La Tene samples from Czechia/Slovakia/Hungary are remarkably diverse. It's quite striking. They seem to be caught in flagrante mixing with what I presume are (southern) Germanic Bronze Age people"

Not just _southern_ Germanic-like. Take a look at how (Hallstatt) I17607 and (LaTene) I20509 plot. They're up there in the CWC/Dutch Beaker zone.

Rich S. said...

Wee e said...

"No: the implication being that THIS is a trivial difference. Not at all one that gets to the heart of the differences. Not one that is at all informative, either. . ."

No, it was not implied that the change from Q to P was trivial; it was stated outright. The implication - i.e., the thing that was implied - was that because that was a relatively minor sound change (at least in the opinion of the guy who said it), therefore the difference between Q-Celtic and P-Celtic was unimportant or trivial.

I merely pointed out that the monikers "Q-Celtic" and "P-Celtic" were chosen as TITLES to characterize branches of Celtic that actually differ from each other in significant ways. Other differences might have been chosen to form the names of the branches. The mere names of the branches aren't complete concatenations of their differences.

Okay, enough of that. My point is that I think it worth considering that the arrivals from the Continent who carried increased EEF might reflect the arrival of carriers of the P-Celtic language innovation.

It might be the case that Q-Celtic was already present in the Isles before the newcomers arrived.

andrew said...

"the paper linked to says that “Between 1000 and 875 BC, EEF ancestry increased in southern Britain (England and Wales) but not northern Britain (Scotland) due to incorporation of migrants who arrived at this time and over previous centuries,"

This is attested in archeology as the time frame of Celtic expansion. It should have been there. The Bronze Age EEF increases mentioned by Davidski are another matter.

andrew said...


"Rob said...

@ Andrew

''Presumably, the language replaced was part of a language family associated with the mostly Anatolian Neolithic LBK migration of first farmers

Sure this is a hypothetical side-comment, but i would still point out that this is a rather impossible scenario given that LBK never reached Britain. British farmers arrived 1,000 years after LBK ended and were vastly different."

Fair point. I stand corrected. Mediterranean route Iberian first farmers (Cardial Pottery roots, I presume). But, basic conclusion: probably spoke the same language family as everyone else in that wave of first farmer migration along the Mediterranean. This was probably a different language family than LBK, but both were probably rooted in a different branch the same Western Anatolian macro-language family as LBK but a phylogenetic level or two up from it, as Western Anatolia is at the root of both the Mediterranean and Danubian first Neolithic expansions.

andrew said...

@Cobra "Now about the Anglo-Saxons and the later Germanic waves we need a specific study about it and how much impact they had." Pretty sure that this has been done. Stephen Leslie, et al., "The fine-scale genetic structure of the British population" 519 Nature 309-314 (March 18, 2015).

Samuel Andrews said...


Genetically, Bell beaker & R1b P312 originated in Western Europe. The most common theory is they originated in the Netherlands which is in Western Europe.

The R1b P312, Beaker folk in Germany, Bohemia, Poland, Hungary came there from the west.

They were apart of a western European family.

Davidski said...


The Bronze Age EEF increases mentioned by Davidski are another matter.


The Early Iron Age in Britain starts in 800 BC.

So the increase in EEF between 1000 and 875 BC obviously happened during the Late Bronze Age.

Romulus said...

In Madison Grant's 1916 book he stated that the Celts arrived in Britain in 800 B.C.

Pretty impressive

Bob said...

@ Davidski

No longer does the EIA in Britain start 800 BCE???
(as a point of detail, the LBA-EIA transition began 800 BCE and the EIA 600 BCE according to most current chronologies. The start of the transition was defined by the appearance of decorated ware pottery).

Congratulations in identifying the new EIA population. I see this as a direct result of the LBA migration with increased EEF - but the origins/routes into England may be more complex than the paper suggests.

I think we can be the first to claim - here - to have pushed back the Early Iron Age transition date to 950-900 BCE. This coincides with the appearance of the first decorated ware pottery at Cliffs End (according to Waddington et al 2019) - but I do need to check this in the Cliffs End monograph. We can also say that also coincides with the arrival of people who created a distinct population in parts of southern England.

I am now pretty sure that I can see this in admixture (so far been looking at English MBA v - Knoviz v Grand Est v EEF v Steppe as sources - not only across the EIA samples - but carrying on into the MIA with differences in individual sites.

As I am new to this, I could use some feedback on whether the tests I am trying are useful and valid. Unfortunately out tomorrow and trying to avoid the inevitable Omicron.

Bob said...

A word of caution on the EIA impact on language.

In looking at the admix of the East Yorkshire 'Arras Culture' sites, the samples are mostly close to English MBA samples. The isotope data does not prove migration/movement of people within these groups. On the flip side, even many archaeologists who do not believe in migrations, believe there may have been a significant MIA movement of people into East Yorkshire from France. Also, the Y DNA is certainly more diverse - and includes R1b-U152 and other 'exotic haplogroups (one commentator on Anthrogenica described it as an A-Z of Celtic Y DNA - possibly overstating it slightly!).

A long winded way of saying, there was probably rather more extensive migration in the Middle and Late Iron Age that is currently very difficult - or even impossible - to identify through standard autosomal analysis(?).

It may be that these later migrations/movements of people, may be very significant in the distribution of Celtic language forms/dialects. Also in the distribution of archaeology features (e.g. hillforts?) and social/economic activities.

I think that this large data set gives us an opportunity to look at the DNA on a more local level - even down to individuals. And to tie this to the site archaeology, including the limited isotope studies that are available.

Davidski said...


More precisely, the start of the EIA in the British paper is defined as ~800 BCE, because all of the 800-400 BCE samples are labeled EIA.

I'm going by the chronology defined in the paper, because otherwise things get very confusing for the readers here.

But if we say that 800-600 BCE is the transition period, then that could mean that the genetic link between Britain and France that I'm seeing dates to the transition period, rather than the EIA. However, I'd need to check the dates of the relevant samples, just in case they date to 600-400 BCE.

Rob said...

This is a good image of 'archaeological cultures' in France in Halstatt period (~ 800 bc ->)

From that article - ''The French regions south of the western Hallstatt province were shared between two Northwestern Mediterranean cultural groups: Ibero-Languedocian Cultures, well documented by numerous huge urnfield cemeteries and some fortified settlements, and Ligurian Cultures, of which less is known. Western and Northern France had different cultural traditions: Aquitania, with its cremations deposited in barrows, shared affinities
with Northern Iberia, whereas Central Western France reveals hybrid influences, originating in the Hallstatt and the Atlantic regions. From Gironde to Flanders, Northwestern and Northern France belonged to the same cultural network and show many relationships with Southern England: all these linked archaeological cultures are called ‘Medio-Atlantic,’ because geographically they occupied a middle place in the networks that for a long time had connected the European Atlantic facade (Cunliffe 2001; Milcent 2006).''

Davidski said...

Admixture analyses won't be able to pick up fine scale differences within Britain and nearby France, because they focus on older components.

Looking at my PCA, it seems that a large proportion of the England EIA samples are quite distinct from the England MBA samples.

CHG Chad said...

@ Samuel

Again...I am speaking about autosomal here.You like or not Latins,Romans,Italics and Iron Age Italy in general were not western Europeans by genetics.Eveybody knows that proto-Italics were people with IE(Beaker) ancestry but when they ended up in the Italian peninsula they mixed heavily with the natives who were mostly a mix of EEF/WHG.Thus, they become more southern shifted.Iron Age Italy(autosomal)has high EEF these people can be considered western europeans? Ydna is also important but when it comes to autosomal DNA these historical populations had little to do with western european genetics.

The western European genetic spectrum has to do with the Celto-Germanic ethnicities around the low countries,Northern France,South Netherlands,Western Germany and also maybe southern England.

Northwest European genetics can be considered ethnicities like the Irish,Northern British,Scottish,Welsh,North Dutch and some folks from Norway and Iceland.People with ancestry from Beakers and Germanics(NorthSea/Scandinavian).

I don't know why you trying to make Italics,Latins,Etruscans western europeans.Everybody knows they werent.If you mean they had R1b western lineages...then yes you are right.But their autosomal especially during IA times and later had little to do with western europe.They would had been closer to Iberia and maybe Austria/Swissland.

LivoniaG said...

@ George - thank you

Burial practices in iron age Britain:

Wow. What a mess. So many of the listings are not reliable.
I assume this is an old list and Reich and Patterson
worked with data better than this.

LivoniaG said...

KM wrote:
"The "trade contacts" identified by archaeologists and similarities in art across the swathe of land linking Denmark and Greece might reflect this. I wonder to what extent this was generated by some process of empire and conquest."

More likely a network of long traders and merchants who could afford to hire, train and support mercenaries as protectors of roads and rivers. A separate operation from local tribal bands, who might otherwise try to get a piece of the action. The groups called the Bastarnae for a long time in Greek and Roman Lit -- with the name coming from "bast" (rope) and "bastagna" (baggage,freight handlers) -- before the word meant bastards --- supplied both wagons and boats plus military protection along their trade routes, which they may have monopolized.

Bob said...

@ Davidski
OK. I am trying to look at sites and distance from other LBA / IA populations (not actual sources because they are not available - but closest proxies that are likely currently available).

If you think that will likely not work, I will try the key site comparisons in NW Europe and also in Celtic Germanic PCAs to see if they show the differences.

Bob said...

@ Davidski
The EIA , and distinguishing 800 to 600 BCE and comparisons of EIA 'sites' may be impossible to resolve. There are so few samples available for the period. Some are just fragments of bone from middens - there may have been excarnation practices in many areas and possibly cremation in other locations. Unfortunately, there are not nice cemetery sites to look at in England for this particular period. It makes an interesting and difficult challenge.

And then the radiocarbon plateau is a huge problem through the Iron Age, so dating the few samples we have may be very imprecise.

Davidski said...

Well, this paper includes a decent number of ~800-400 BC English samples, and, unlike Bronze Age English samples, a large proportion of them do overlap with similarly dated French samples in my PCA.

So this might be relevant to the Celtic expansions, and indeed perhaps more so than the rising level of EEF in Britain throughout the Bronze Age.

But yeah, more Early Iron Age British samples, from more precisely dated sites, would certainly be helpful.

Rob said...

I noticed, if we add England_LBA # high EEF into the same North Euro PCA, then they overlap with French Samples too
Without appropriate French MBA & LBA it is difficult to exclude counter-current movements from England into France.
I'm inclined to view the genesis of Celtic as a complex process, possibly longue duree

Davidski said...

I ignored the high EEF individuals because they exist in all of the time periods, including the EBA and EIA, and it's impossible to know what they represent.

For instance, two out of three of the high EEF England LBA samples actually still look quite British despite their high EEF.

So they might represent different things for the different and same time periods, like both high EEF survival and sporadic immigration from somewhere in France or even Iberia.

In my opinion, the fact that the mainstream EIA samples overlap strongly with the French samples is the really interesting thing here.

Slumbery said...


The North Euro PCA overlap between England EIA and French IA looks good only by the colored areas, but the sample overlapping itself is a bit muddier. The two French samples, that are most responsible for the area overlap, also overlap with England LBA and MBA very well. It is possible that those two represent a movement into the opposite direction. And if we go with that assumption and remove them from the equation, then a good chunk of the area overlap between England EIA and French IA disappears.

At the other hand I have to admit that some of the England EIA samples are placed where a French IA + England BA mixed population would be.

A said...

@ Cobra,

IA Romans had 50% + ancestry from a Northwest European-like source, but according to you 'they had little to do with western european genetics'. Amazing.

Rob said...

I think you can model 'regular' England IA as LBA + LBA/high EEF
But the Yorskshire guys need Central European admixture (e.g. Knoviz).

Matt said...

I've tried a few things using G25 to sort of control out the differences in EEF/WHG/Steppe and see if there is any residual difference remaining in the type of EEF between IA vs pre-IA. Using multivariate regression to try and isolate a different type of EEF. I didn't get too much of a signal to be honest - mostly just came out as a Globular Amphora like thing.

This kind of contrasts with a tenative claim made by Nick Patterson in the comments that there is a clear difference seen in their qpAdm models between the EEF IA ancestry source in IA and CA_EBA.

(See here for the comments:

One other idea I've thought about is trying to get three sets of ancient British samples who should have a similar average steppe proportion (by taking those with higher than average for some eras and lower than average for others), and then comparing them directly with formal stats. Any difference remaining should be due to the drift+differences in EEF ancestry?
Set like this, where all should average close to 55% Yamnaya in a G25 Vahaduo model is:

1: CA_EBA (10 samples): I5379, I2597, I2445, I3255, I14543, I1770, I2601, I18606, I5377, I1775 (average 682450 SNPs, varying from 404131 to 860982).

2: MBA_LBA (19 samples): I7570, I7573, I19859, I7568, I19860, I7572, I7640, I7571, I2574, I7574, I19913, I7626, I2448, I7577, I7575, I7580, I16488, I7628, I13713 (average 673430 SNPs, varying from 315701 to 821054).

3: IA (20 samples): I11154, I16441, I19211, I20586, I5507, I14804, I12779, I20623, I13687, I14348, I12790, I5505, I5508, I14104, I20621, I13616, I14351, I14097, I0160, I14096 (average 734036 SNPs, varying from 414550 to 1142730).

Can't totally control for all the variance (although I tried, I did get variation from 56.8% to 55.4% still, about 1.4%) but it should be close enough to hopefully lens out the EEF/WHG part in formal stats.

In theory then, if the Yamnaya ancestry is the same, then any remaining differences in ancestry would be in the non-Yamnaya part. So they should show up in f4 statistics like f4(1,2;Globular_Amphora, Sardinia_MN) and so forth. If Patterson is correct this should particularly be the case comparing Sardinian:Balkan farmers (e.g. Cardial vs Danubian). The trouble is that this subset might be challenged as unrepresentative...

(Quick illustration of how close the average of these subsets are in PCA and Vahaduo admixture relative to the unselected sets:

If we look at the samples from England, it does look like the change in ancestry over time is associated with a declining ratio of WHG:EEF -

Romulus said...

In Bohemia as per the recent paper on that, the arrival of the Unetice Culture saw a ~50% Autosomal replacement and a rise in EEF ancestry. Surprise surprise as the Bronze age reaches Britain(from Unetice derived central European cultures) we see an equivalent change. The only major difference is that instead of I2 its R1b-U152. I2 were the smiths and the U152 the traders?

Samuel Andrews said...

"The western European genetic spectrum has to do with the Celto-Germanic ethnicities"

Italic is a sister language to Celtic. If Celtic is West European, how is Italic not West European?

There's no rule that says only northwest Europeans can be West European. There's also Southwest European with higher EEF ancestry. This is was Italic tribes were.

CHG Chad said...

Distance to: ITA_Rome_Latini_IA
0.04183473 Spanish_La_Rioja
0.04570194 French_Corsica
0.04771550 Spanish_Castello
0.04843403 Spanish_Baleares
0.04868291 Spanish_Menorca
0.05019594 Spanish_Pirineu
0.05039758 Spanish_Murcia
0.05054424 Spanish_Castilla_La_Mancha
0.05079895 Spanish_Aragon
0.05084403 Spanish_Soria
0.05091671 Spanish_Terres_de_l'Ebre
0.05102282 Spanish_Girona
0.05106171 Spanish_Catalunya_Central
0.05119754 Spanish_Lleida
0.05153013 Spanish_Mallorca
0.05171704 Spanish_Peri-Barcelona
0.05218196 Spanish_Valencia
0.05253162 Italian_Bergamo
0.05257496 Spanish_Alacant
0.05258964 Spanish_Eivissa
0.05296461 Spanish_Cantabria
0.05298299 Spanish_Cataluna
0.05301040 French_South
0.05345731 Spanish_Camp_de_Tarragona
0.05348050 Spanish_Penedes

Distance to: ITA_Prenestini_tribe_IA
0.01431569 French_South
0.01729684 Spanish_Aragon
0.01775701 Spanish_Pirineu
0.01786436 Spanish_Lleida
0.01813429 Spanish_Castello
0.01824498 Spanish_Cantabria
0.01846337 Spanish_Navarra
0.01914123 Spanish_Catalunya_Central
0.01970117 Spanish_Terres_de_l'Ebre
0.02013640 Spanish_Penedes
0.02051695 Spanish_Girona
0.02094324 Spanish_Peri-Barcelona
0.02099927 Spanish_Barcelones
0.02109724 Spanish_Valencia
0.02184882 Spanish_Pais_Vasco
0.02226555 Spanish_Camp_de_Tarragona
0.02300836 Spanish_Cataluna
0.02314624 Spanish_Castilla_La_Mancha
0.02399636 Basque_Spanish
0.02414738 Basque_French
0.02506561 French_Auvergne
0.02558435 French_Occitanie
0.02593674 Spanish_Alacant
0.02633019 Spanish_Mallorca
0.02710038 Spanish_Baleares

Distance to: ITA_Proto-Villanovan
0.02645996 Italian_Lombardy
0.02887154 Italian_Piedmont
0.03195847 Italian_Bergamo
0.03280277 Greek_Thessaly
0.03301222 Italian_Trentino-Alto-Adige
0.03311316 Italian_Liguria
0.03393619 Italian_Veneto
0.03438140 Swiss_Italian
0.03587732 Italian_Tuscany
0.03599655 Greek_Macedonia
0.03609949 Albanian
0.03659517 Italian_Northeast
0.03887975 Italian_Marche
0.04005437 Greek_Central_Macedonia
0.04055784 French_Corsica
0.04266896 Gagauz
0.04297156 Italian_Umbria
0.04306784 Bulgarian
0.04329836 Rumelia_East
0.04357756 Spanish_Menorca
0.04381777 French_Provence
0.04417170 Greek_Peloponnese
0.04427875 Spanish_Eivissa
0.04501181 Spanish_Mallorca
0.04542032 Spanish_Baleares

Distance to: ITA_Boville_Ernica_IA
0.03638036 Spanish_Terres_de_l'Ebre
0.03700614 Italian_Lombardy
0.03752470 Spanish_Castilla_La_Mancha
0.03825974 Spanish_Valencia
0.03838833 Italian_Bergamo
0.03915744 Spanish_Aragon
0.03923277 Spanish_Eivissa
0.03935741 Spanish_Andalucia
0.03940448 Spanish_Menorca
0.04065794 Spanish_Navarra
0.04094360 Spanish_Pirineu
0.04099104 Spanish_Lleida
0.04109052 Spanish_Peri-Barcelona
0.04142351 Spanish_Asturias
0.04195715 French_Corsica
0.04210027 Spanish_Barcelones
0.04292418 Spanish_Alacant
0.04293783 Spanish_Cantabria
0.04307153 Spanish_Murcia
0.04352119 Italian_Trentino-Alto-Adige
0.04380708 Spanish_Camp_de_Tarragona
0.04406153 Spanish_Baleares
0.04410448 Spanish_Castello
0.04445591 Spanish_Mallorca
0.04452170 Spanish_Cataluna

Distance to: ITA_Etruscan
0.02141200 Italian_Bergamo
0.02202841 Italian_Lombardy
0.02547965 Italian_Trentino-Alto-Adige
0.02582660 Spanish_Menorca
0.02742894 French_Corsica
0.02758314 Italian_Veneto
0.02960911 Spanish_Baleares
0.03005067 Spanish_Terres_de_l'Ebre
0.03115347 Italian_Piedmont
0.03147144 Spanish_Eivissa
0.03173586 Spanish_Castilla_La_Mancha
0.03178110 Spanish_Mallorca
0.03197242 Spanish_Peri-Barcelona
0.03258140 Italian_Liguria
0.03295674 Spanish_Lleida
0.03397964 Spanish_Girona
0.03423275 Spanish_Catalunya_Central
0.03441671 Spanish_Valencia
0.03455966 Spanish_Aragon
0.03497162 Spanish_Camp_de_Tarragona
0.03503611 Spanish_Murcia
0.03509346 Spanish_Barcelones
0.03530180 Spanish_Pirineu
0.03621935 Spanish_Penedes
0.03637690 Spanish_Castello

CHG Chad said...

@ Samuel

They were southwest indeed.They had high EEF mix and adittional steppe/WHG admixture.But you can't say they were proper western europeans like the French or the Belgians for example.They were in the cline of North Italy-Iberia.Btw,what the language has to do with genetics and autosomal DNA?

Mycenaeans(bronze age Greeks) were IE speakers.Their steppe admixture was not more than 20%...does it making them Northeast or Eastern europeans?You can say that Mycenaeans were like the Polish or the Ukranians?

Wastrel said...

I hope people don't mind me trying to inject a little bit of clarity on some points...

- the Bell Beaker settlers of Britain absolutely, emphatically, did not speak Celtic languages or their ancestor

- AIUI, Celtic is assumed to have arrived ~800-600ish. It's assumed to have been dominant by 600ish, when the name 'Albion' is attested, which looks Celtic. This period also (according to Davidski) looks promising genetically, matches up with an influx of continental cultural featues in the archaeology, and makes sense linguistically.

- if that were pushed to 1000-800ish, I'm not sure it would absolutely break anything major, it just wouldn't fit as well linguistically, archaeologically or genetically.

- we know there was at least one later invasion, the Belgae, late in the 1st millennium. There may well have been multiple waves of migration (Irish mythology certainly thinks so!)

- cultural contact was extensive. In Caesar's time, for instance, exiled Gauls could hop across to Britain and start up their own little clans there, and Belgian kingdoms could speak of controlling some land in Britain.

- it's debated, but "P-Celtic" looks real. It's hard to know when it happened - did both groups move to Britain at once, or did P-Celtic come later?

- however, P-Celtic/Q-Celtic doesn't look important. It's only about half a dozen minor features. There's no reason to think the Brythonic/Goidelic groups were wildly separated and arrived at totally different times - closely-related neighbours seems more likely.

- we even have the same tribal name in both forms (Pretani/Cruthin), implying that the split wasn't seen as a major issue - a split within and across tribes rather than strictly between them

- P-Celtic was in Britain by 300BC at the latest, when "Pretanic" is attested. It spread across Britain - undoubtedly as far as southwestern Scotland, and probably across the whole island.

- "Insular Celtic" is a much bigger deal, a really major divergence from the continent (but could easily be areal influence rather than shared origin).

- obvious hypothesis is that P- and Q-Celtic groups came to Britain 800-600, either simultaneously or in close succession. They remained in contact with continent, and their languages didn't diverge much phonologically, but they WERE living together, and separate from Gaul, for long enough to develop some REALLY WEIRD syntactic oddities not seen on the continent.

- it's probably not wildly wrong to say that Brythonic vs Gaulish in Caesar's time was something like English vs German in, say, 1600: many recognisable words and not a hard language to learn, but also some major differences in the grammar. [as for Q vs P: note that Low German is genetically closer to English than to High German, but sprachbund effects make it look much more like High German than English. Likewise, Brythonic was probably genetically closer to Gaulish, but looked a lot more like Goidelic due to areal convergence].

Romulus said...


"This is was Italic tribes were."

I fully agree.

Wastrel said...

Answering Wee e's questions:

In 1kbce, most of France probably spoke languages with no current ancestor. They would probably have been languages brought by steppe/CWC invaders, and parsimony suggests that they were IE languages, because we know that that invasion was associated with IE languages, and there's no evidence of it being associated with any other language family. [all the languages strongly associated with these population movements, now and historically, are all IE languages]. However, this is not certain. It's entirely possible that some other related or even unrelated language was also spoken by some immigrants - Basque and Etruscan were both spoken by people with steppe-like genes, and while it's assumed that they were relict languages (the language survived even as the genetics changed), we don't know this for sure.

However, there's no good evidence of widespread non-IE placenames in most of France, which again makes the IE assumption more attractive.

Regarding your three suggestions:

- Celtic in a strict sense probably didn't exist yet. However, Proto-Celtic was probably spoken somewhere at that point. 'Somewhere' could have included parts of eastern France. Or maybe even all of France, we don't know. "Italo-Celtic", "Para-Celtic" or "Para-Italo-Celtic" are all good guesses for languages in France at that time.

- Germanic definitely didn't exist. Proto-Germanic is much, much younger than Proto-Celtic. Proto-Germanic shows signs suggestive of early contact with Balto-Slavic; however, it also shows some non-IE (LBK?) loanwords in common with Italo-Celtic. So they would have been spoken in similar areas. In particular, we know that Proto-Germanic was spoken in contact with Celts, because there are many Celtic loanwords into Germanic, particularly relating to war and politics; Celts are assumed to have been in some way dominant over Pre-Proto-Germans (not necessarily as rulers, perhaps just as violent neighbours). Could a language closer to Germanic than to Celtic have been spoken in some areas that were later Celtic, eg northern France? Absolutely! But there's no reason to assume this.

- DO NOT FORGET that IE languages didn't emerge fully formed and distinct one day. They diverged over time. In 1000BC, they had diverged much less. The differences between dialects would have been smaller; and in particular, before the Celtic expansions it's likely that there would have been a continuum of dialects in western Europe. If we did have records of a language from Belgium circa 1000BC, it's not clear that we would be able to say which IE branch it was closer to - it would probably be in between. It's certainly possible that one reason Germanic looks a bit odd grammatically is that all the intermediate forms in northern europe were extinguished by the Celtic - and later the Germanic - expansions.

- Bear in mind: Germanic is the ONLY known survivor of ANY language spoken in Western or Northern Europe prior to 800BCish (assuming Italo-Celtic is from 'Central' Europe), other than Basque and (some fragments of) Etruscan - and Germanic is attested very late compared to Celtic and Italic. This is a huge area about which we know nothing.

- regarding Basque: we know that Vasconic was spoken in much of southwestern France in Caesar's time. We don't know whether the Aquitanians were an expansion out of Iberia, or the Basques were an Aquitanian expansion into Iberia (certainly much of the Basque country wasn't Basque until recently). It's possible that the Aquitanian territory was much larger before the Celts came - but there's also no evidence that this is the case, so it's usually assumed that it wasn't.

Wastrel said...

- on the specific issue of Pictish: placenames, personal names and loans into Scottish Gaelic all indicate the presence of some P-Celtic language throughout Scotland, which we can call Pictish. Regarding the ogham: caution is advised. These were late inscriptions, made in a Pictland that was heavily Gaelicised, and in direct emulation of Gaelic ogham; many of the inscriptions are hard to interpret. It's certainly not impossible they may already have borrowed the word 'maqi', particularly in the context of an inscription formula. We also don't know their spelling conventions: given that they would have known that the Irish used 'Q' to spell a sound that in Pictish was 'P', they may simply have used 'Q' to spell that sound themselves, particularly in words that they knew were cognate to Irish. In any case, I'd put more emphasis on the local names in the inscriptions than on the imitated inscriptional formulae.

Romulus said...

There is a sample on yfull from this branch of I2 from Pakistan.

1500-1250 BCE Czech_MBA_Tumulus I-L1229 I2a1b1a1b1
1300-800 BCE Czech_LBA_Knoviz I-L1229 I2a1b1a1b1
1000-800 BCE Czech_LBA_Knoviz I-Z2069 I2a1b1a1b1a
750-415 calBCE Slovenia_EIA I-Y3672 I2a1b1a1b1a1a~
800-550 BCE Czech_IA_Hallstatt I-S18331 I2a1b1a1b1b

Rob said...

@ Wastrel

''However, there's no good evidence of widespread non-IE placenames in most of France, which again makes the IE assumption more attractive.''

Not that I necessarily disagree, however names & words in general might not be the best way to uncover language layers because when speakers shift languages the major impact is in morphosyntax & phonology rather than lexis. Lexical borrowing and intrusions occur most in a steady-state when societies keep their independent languages , e.g classic Balkan sprachbund. I dont see why non-IE weren't spoken by some of the creolized, post-Beaker societies south of the Seine (as you did mention)

As a counter-example, linguists have often remarked that there is a paucity of Celtic/ Brythonic placenames in England, but we know such languages were spoken there before 450 AD

Andrzejewski said...

@Rob “ As a counter-example, linguists have often remarked that there is a paucity of Celtic/ Brythonic placenames in England, but we know such languages were spoken there before 450 AD”

The paucity of Brythonic place names in England correlates to its paucity in the English gene pool: 60% -80% of which is Anglo-Saxon.

Andrzejewski said...

@Wastrel “ Proto-Germanic shows signs suggestive of early contact with Balto-Slavic; however, it also shows some non-IE (LBK?) loanwords in common with Italo-Celtic.”

How do you know that?

LBK may’ve been affiliated with Ötzi’s language, or maybe even with Tyrrhenian languages like Etruscan, but no way did it contribute to Germanic; if anything then GAC and TRB are closer genetically to the Cardial Pottery.

And Celts have had much more EEF than Germanics, therefore the false claim to any substrate theory doesn’t hold water, because Celts had no substrate whatsoever.

Davidski said...

There's only about ~30% Anglo-Saxon ancestry in England overall. Probably more than that in the east, but much less than that in the southwest.

Rob said...

@ Andrze

'The paucity of Brythonic place names in England correlates to its paucity in the English gene pool: 60% -80% of which is Anglo-Saxon.''

That doesn;t ring true, although there might be 5th cc. A-S immigrants who are 80% continental Germanic , or even more. But if you took Mercians or Saxons from 8th cc, the average Brythonic ancestry would be more than 30%

LivoniaG said...

Wastrel said...
"the Bell Beaker settlers of Britain absolutely, emphatically, did not speak Celtic languages or their ancestor"

Wastrel - Can I ask how you know that with such certainty?

It's possible the different Bell Beaker types spoke different languages in Holland, Iberia, Sardinia, Sicily and Poland. We have no written records of any languages in most of those places till more than 2000 years later

Terrific paper on the intensive bronze mining and production at the great Orme in Whales that took off about 1600-1400BCE and was one of the earliest producers of true bronze in Europe,with chemical proof it reached Holland and Sweden in finished form.
In the paper they mention that know how and experts might have come in from the earlier copper works at Ross Island of Ireland which seems to be a Beaker colony.
Whatever language or languages they were speaking at the Great Orme Mine and along the extensive network that moved the bronze rapiers and other heavy objects throughout middle and southern Britain and beyond,it would have been handy for the 1000s of people involved to speak the same language. This all happened 600 years before the comming of the iron age.

Ric Hern said...

So how did Ireland become Celtic speaking ? Is there any genetic link between the EEF rich people or the early Iron Age and Ireland ?

We already know there was a link to Unetice with the Rathlin samples but they were basically contemporary with the early Unetice several thousand kms to their East...? We also see some Archaeological similarities during the Middle Bronze Age from Ireland to as far East as the Wietenburg Culture in Romania. We know that Bell Beaker, Unetice and Urnfield influenced vast areas culturally and each could have had a dialect Levelling effect throughout the areas where Indo-European was spoken. If I remember correctly there are very little evidence of Urnfield and later Hallstatt Cultural impact on Ireland. So this only leave us with Bell Beaker, Unetice and La Tene....

Garvan said...

This is a quote from one of the co-author’s, Barry Cunliffe, "If the Mid-Bronze Age move had any effect at all on language, then the simplest hypothesis would be to see it as a vector for introducing, or strengthening Brythonic, if so, then Goidelic had to have arrived earlier, either with Beakers, or earlier."

In my view, Cunliffe likes to take positions which challenge orthodox views, to generate constructive debate. But I am intrigued to know what he meant by “or earlier” in the above quote.

Rob said...

Some British scholars had the view that Celtic developed as early as the Neolithic, with megalithic groups. The modified view was BB period, after the early adna studies came out. Even then, the Neolithic could be hitch-hiked because of the idea that BB diffused ex Iberia

Dranoel said...

Hi to all.
Again, I would just like to make a small comment about the Z2103. Both the main article about the UK and the article quoted by Romulus mention the Z2103 (although the one in the Netherlands is not sure yet, it shows a certain trend anyway). In addition, we know that Z2103 was in the CWC of the Czech Republic and BB in Poland, Hungary. This is further evidence that we can no longer speak of Z2103 as a typically Eastern SNP. As I wrote before, both the archaeological DNA research and the modern distribution of FTDNA collections indicate and confirm that a certain group of Z2103 migrated with others to NW Europe. Would it be a big surprise to find the Z2103 in a Single Grave? In my opinion no.

In previous posts I have suggested that the Z2103 in NW Europe may have spread further as a component of Unetice or Tumulus migration. This would also fit in with the present day results in FTDNA (although I take this only as an uncertain curiosity). Earlier I wrote that it seems to me that Z2103 accompanied the Germanic expansion / co-created these cultures, mainly in the eastern part of their range). If the sample from the Netherlands that was adjacent to another U106 turned out to be true, that would be another good argument. In the Netherlands we also have the Elp (Tumulus influenced) and Hilversum cultures, which are the so-called Northwestblock - presumed, some kind of intermediary between Germanic and Celtic. Maybe they "moved" the Z2103 to both the Northeast and the British Isles?

Regular readers of Eurogenes probably associate Davidski's map with the "Bell Beaker highway" - it will turn out that in a minority, but we have a Z2103 on every "episode" of it :)

Rob said...

From the study - ''Later individuals in North-Central Europe have similar EEF proportions,
consistent with substantial continuity through the LBA-IA''

not so sure about that. For ex; la Tene individuals from Czech R. don;t look the same as Hallstatt or Knoviz.

Ric Hern said...

If there were continuity in Northern France from the Early Bronze Age then which route did the EEF heavy people take to Britain ?

Davidski said...

Right, I think the biggest problem with this paper is the suggestion that the EEF-heavy people are directly relevant to Celtic expansions.

This is unlikely, because there was no coherent EEF-heavy population that moved into Britain, and you really need a population to move there to change the language, not just some outliers.

Steppe said...

How high is the proportion of the steppes / CWC / BB among the Celts from continental Europe?

alex said...

"Samuel Andrews said...

Genetically, Bell beaker & R1b P312 originated in Western Europe. The most common theory is they originated in the Netherlands which is in Western Europe.

The R1b P312, Beaker folk in Germany, Bohemia, Poland, Hungary came there from the west.

They were apart of a western European family."

Yes, except the P312 people in the Netherlands didn't speak Latin or Etruscan. Cobra kindly pointed out to you that Latins and Etruscans weren't very northern autosomally. Etruscans were 63% EEF, more than a lot of modern southern Europeans. Anyway, if they weren't Mediterranean we would probably know very little about them and their culture so I guess it's a good thing they were Mediterranean.

Romulus said...


"Celts are assumed to have been in some way dominant over Pre-Proto-Germans (not necessarily as rulers, perhaps just as violent neighbours)."

Germanics were dominating Celts out of the gate and it's the core reason Caesar had to intervene and conquer Gaul. What you stated is a falsehood and coping fantasy.

Samuel Andrews said...


Nice catch. L1229 has popped up many times in the Bronze age so far.

The Tollense warrior I2a is also L1229. As is most of Unetice I2a and one of the I2a in Bronze age Serbia.

L1229 must have expanded in a major way in the Bronze age.

I haven't looked at the Celtic data at all. Is their I2a, mostly L1229?

Samuel Andrews said...


Are you thinking Neolithic farmer rich people lived as an underclass in Indo European societies?

Hence, the EEF-rich outliers in Bronze age and Iron age Northwest Europe.

Samuel Andrews said...

I haven't looked at the Celtic DNA yet. I'm too busy making a video the Bohemian ancient DNA study.

Are there just sporadic EEF rich individuals in Bronze age Britain? With a majority of Bell Beaker-like individuals? If so, then yeah that's not a population movement.

weure said...

This paper has also made clear that the so called Elp culture (North Dutch/NW Germany) is in fact founded by a Tumulus warrior elite.

This can be show by a specific subclade of R1b U106 namely Z156 and more specific Z304.The Dutch BA/IA samples are relative small, but still deliver 'a bunch' of Z156 > especially Z304 (>DF96/98). As does Bohemia.

I7196 DF98>S1911(1 read)>S1894(1 read) NextSeq 500 2200-1950 BCE 40+ Older Únětice Jinonice, Prague, Czechia
I4070 U106>>Z381(3 reads)/ (?Z304- 1 read) NextSeq 500 1880-1657 BCE 26-49 Early Bronze Age, Elp De Tuithoorn, Oostwoud, Noord-Holland
I11972 U106>>Z381 (-Z301) 1240K 1501-1310 calBCE 50-65 Middle Bronze Age Westwoud, Binnenwijzend, Noord-Holland
I17019 U106>>Z381>Z156>>Z306>>Z304 1240K 1421-1216 calBCE Middle Bronze Age Vlaardingen-Krabbeplas, Zuid Holland
I13788 U106>>Z381>Z156>>Z306>Z304 1240K 1300-800 BCE >50 Late Bronze Age Chouč, Hrobčice, Teplice, NW Bohemia
I23978 U106>>Z381>Z156>S5520>FT221936 1240K 742-400 calBCE infant Early Iron Age Zagorje ob Savi, Upper Camiola, Slovenia
I15950 U106>>Z156>>Z304>BY12480>BY12482/Y28944 1240K 480-390 BCE 35-60 La Tène Teplice, Ústí nad Labem , NW Bohemia
I12907 U106>>Z381>>Z306>>Z304 1240K 356-57 BC c8 Iron Age Uitgeest-Dorregeest, Noord-Holland.
(from collection of Ray Wing)

Of course we can't state that ALL Z304 spread were carried by the Tumulus culture (we never can because would mean a 100% overview we will never have....)..

But we can state that the Elp samples (= Tumulus along the North Sea Coast) in the Netherlands that are connected with R1b U106 are all, without exception Z156, and most of them Z304.

Of course there are other at stake R1b 312> DF98 (>subclades) for example are also present in the Dutch MBA samples.

The archeological context shows a specific relationship with the Fulda Werra culture, one of the dignified Tumulus culture of middle Germany, connected via the Weser (Werra is Weser) to the North Sea. And the Rhine most probably also played a part Rhine Main group.

This picture (original by archeologist Bergerbrant) was added by me a few weeks ago, now after this paper it has more solid ground!

I'm convinced (and this paper gives ground to it) the Elp culture is a Tumulus offshoot via 'middle Germany' and was originally rooted in Unetice/ Central Europe (Bohemia and surrounding area's). It's most probably also connected to the Nordwestblock language that had probably "Italo-Celtic" kind of features.

Genos Historia said...


The single U106 sample in Unetice is an immigrant from Western Europe. He has more EEF than all other Unetice samples and clusters with modern Dutch.

Tom said...

Spain is a western European country too. Even Italy could be argued to be a western European country.
The UK.
All western European countries. Germany is the intermediate between east, north, west.
Etruscans and Romans were southwest Europeans.

CHG Chad said...

@ Romulus

True.But the tribes below rhine mixed badly with them and they failed to keep their genetics alive.With exception North Dutch and North Germans the others have assilimated in some way to the Celtic world.

@ Alex

Even modern North Italians are more northern shifted in some way from the Germanic invaders there.While Northeast areas have Slavic DNA like in Austria.

Davidski said...


There are both highEEF and highWHG individuals in Bell Beaker Britain.

Then highEEF individuals persist at least until the Early Iron Age, but I reckon we'll eventually also see them popping up as late as the Middle Ages.

So I think that we're dealing here with a very complex situation of persistent pockets of pre-Beaker genetic variation, as well as sporadic movements of people from various parts of continental Europe.

I doubt that there were population movements of highEEF peoples into Britain at any point after the Neolithic.

And this is my main problem with this paper, which seems to have confused this complex picture with the Celtic expansion.

The Celtic expansion probably happened during the Early Iron Age, that is between 800-600 BCE, and, ironically, had little impact on the British ancient ancestry components. But it did link Britain with France genetically for a short time in terms of fine scale (recent) ancestry.

weure said...

@Genos no evidence he Unetice/Jimonice R1b U106 (2000 BC) was an immigrant, let alone that an enhanced EEF was a sign of NW BB migrant on the contrary!

Romulus said...


What you are suggesting is ruled out in the paper

We considered the
117 possibility that the rise in EEF ancestry in southern Britain was due to a resurgence of
118 archaeologically less visible populations with more ancestry from people living in Britain
119 during the Neolithic. However, our attempts to model IA populations of England and Wales as a
120 mixture of groups who lived in Neolithic and C/EBA Britain failed at high significance
121 (Extended Data Figure 1). This is due to LBA and IA populations in Britain sharing excess
122 alleles with Neolithic populations from continental Europe not seen in earlier groups from
123 Britain (Supplementary Information section 3). The most plausible explanation is migration
124 from outside Britain into southern Britain in the M-LBA.

Davidski said...

The scenario I posited is not ruled out in the paper.

Here's that scenario again...

So I think that we're dealing here with a very complex situation of persistent pockets of pre-Beaker genetic variation, as well as sporadic movements of people from various parts of continental Europe.

Rob said...

IMO this supports a model of gradual Celtic formation. I might put some thoughts together soon

Davidski said...

I don't think it was gradual as far as Britain is concerned.

I just think that a whole bunch of Celts armed with iron migrated to Britain after 800 BCE, and genetically they had a short-lived by significant impact in the south.

After that, Celtic languages quickly diffused north and into Ireland.

Romulus said...


That's fair I just wanted to point out that they did test the idea of a resurgence of local Neolithic ancestry in the paper and came up against it.

All I have to say about Celts is that they were seriously shitty warriors. There is nothing badass about running into battle naked when it results in your slaughter.

10,000 EEF Rich Romans annihilate 230,000 of these British Celts
400 dead 80,000 dead

Romulus said...

A resurgence of local Neolithic ancestry would really help explain what was going on with Stonehenge so it's unfortunate to see that idea put to rest.

Andrzejewski said...

@Samuel Andrews Whst happened to all those WHG rich Hungary BA like samples? What do you think is responsible for their decline?

Andrzejewski said...

@Romulus “ 10,000 EEF Rich Romans annihilate 230,000 of these British Celts
400 dead 80,000 dead”

EEF but also CHC and Levant_N rich Roman by that time.

I’m wondering how come the Romans couldn’t use their advantage against the Celts to win against Germanic tribes?

Rob said...

@ davdiski

Well I certainly don’t think BB spoke “Celtic”, but I don’t see evidence of Celtic warriors invading Britain
These people in Yorkshire look like economic migrants & political refugees from eastern Hallstatt
The EEF rich LBA guys could be from Iberia, c/w with preferred Sardinian N source
Btw gradual doesn’t mean smooth

Romulus said...


"EEF but also CHC and Levant_N rich Roman by that time."

Maybe that was the secret to their victory, the original Romans were defeated by the Celts and Rome was sacked by them back in 400 BCE. Acquisition of Greeks and Greek warfare, or maybe Trojans.

Germanics learned from the Romans just as the Romans learned from the Greeks. The Germanics were consistently defeated by the Romans for centuries before things began to turn around with Arminius, who was raised Roman and served in the legions.

CHG Chad said...


Yeah but they failed to conquer Scotland. They just built a wall for protection.


Isn't it a little bit early for West Asian admix among Roman's? I think the period that Roman's invade Britain is the beginning of imperial times. You think they were already mixed with immigrants from West Asia?

Davidski said...


Here are some of those England EIA samples that look rather continental to me in terms of fine scale recent ancestry.

I14747 514-391 calBCE England_EIA England, Kent, Cliffs End Farm

I19861 779-541 calBCE England_EIA England, Wiltshire, Rowbarrow

I19287 761-422 calBCE England_EIA England, Wiltshire, Amesbury Down

I16600 713-381 calBCE England_EIA England, Wiltshire, Amesbury Down

I20585 800-400 BCE England_EIA England, Oxfordshire, Yarnton

I think these people may have been recent descendants of the first Celtic speakers of Britain.

Andrzejewski said...

@All Can anyone post a summary of the various ancestry composition and subclade variety et week: Bell Beaker, Unetice, Hallstatt, La Tene and Nordic BA?

I read a few years back some source who asserted that the switch (or evolution) from Unetice to Hallstatt involved a Scythians-Sarmatian or Dacian-like ruling elite. I personally think that it’s nothing but rubbish.

Romulus said...


Would you be willing to give your thoughts on this sample? I27379
It's a famous one:
I've not found an autosomal analysis of this one specifically in the paper.

Davidski said...

I27379 looks very continental and thus, in my view, quite the Celt.

Romulus said...

thanks, that would seem to confirm this theory:

Arguably the most elaborately equipped warrior grave ever found in England, archaeologists believe he was a refugee French Gallic fighter who fled Julius Caesar's Roman Army as they swept across continental Europe around 50BC.

weure said...

@Genos Historia in MBA there was a flux from Tumulus elite warriors. Their grave goods/ richt Tumuli were exotic for the (North) Dutch/ NW Germany (see Butler 1986/ Fokkens 1998). They meant the end of BB culture in the NW.

This flux was from Central Europe to the NW. The Dutch BA samples (with mainly R1b Z304) were derived (grave goods) from Tumulus cultures of the middle Rhine (Rhine-Main group) and the Weser (Fulda Werra). Both mounding in the North Sea. They were a kind of proto-Celts.

The Fulda Werra group is from Hessen....from the later on big Celtic Glauberg find.

Copper Axe said...


Thanks for the sanity in your feedback for the article. I was scratching my head looking at it and thinking that their data could very easily be interpreted as supporting an iron age migration of Celts to the isles.

A happy new year in advance! Or maybe it isnt much in advance as Australians live in the future. Hopefully 2022 will be a big year for Eurogenes!

Garvan said...

Thanks to Rob for his response to my question on the co-author’s, Barry Cunliffe quote, and for Davidski on for sharing his views. Generally, I like considering all opinions before forming my own, but a Neolithic origin of Celtic is too much even for me.

Matt said...

I thought it was interesting that the RoH analysis didn't indicate very large changes in population size between LBA to IA, as compared to LBA Vs CA EBA. That seems to suggest that the idea that iron farming tools led to dramatic population increases may not hold up, or that the chronology doesn't accurately reflect their adoption in NW Europe.

Having the power to determine actual changes in human population size will be important. (E.g. people didn't necessarily think steppe offshoot pastoral groups would show a big increase in pop size from 3200 to 2500 BCE, but direct genetics say probably yes. While Iron Age increase in pop size maybe no.)

That's somewhat consistent with, for Britain, archaeobotany - . Although that can pick up spike declines that adna cannot see (or not without even more samples).

Ric Hern said...

Maybe the Celts used the Custom of Heroes fighting each other and the losers tribe then bow the knee without massive bloodshed all around. This could explain how a small military force could have had a major impact.

Ric Hern said...

With a spectacular Hero at the helm the average fighter needn't be very skilled.

Moesan said...

I read here some affirmations based on little, and some interesting posts too. I answer to more than one blog guest.
Knowing when and how languages are transmitted without old writings is a very hard work. What is sure is that these shifts are submitted to certain laws where number, political organisation, cultural and economical level, military power, social organisation (elite exogamy and fosterage, slavery…) and so forth.
What have we, if I rely on some readings ?
1- A majority of Celtic names for Scotland rivers and streams, the most of other water courses names being I-E.
& : WFH Nicolaisen wrote (1976) :
[(a) There are river-names in Scotland which cannot be explained as Celtic but which, although too early to be Germanic, are very likely Indo-European.
(b) Many of these bames are identical with names elsewhere in Britain and on the Continent. While in quite a number of instances complete identity is found, in others common stems are combined with common derivatives morphemes, fitting well into a general pattern.
© There are quite a few names on the fringe of this set which, mostly for phonological reasons, cannot be ascribed with the same certainty to this pre-Celtic Indo-European stratum. Some of these may indeed belong here but cannot be distinguished. Others may be Celtic.
(d) It would be wrong to regard the names in question as art of a ‘system’. The more flexible term ‘network’ seems to be much more appropriate.]
plus [...but predominantly at the idea of a pre-Celtic, pre-Germanic, pre-Italic, pre-Illyrian, pre-Venetic, pre-BalticIndo-European river-nomenclature in Europe, an idea first proposed and subsequently supporte and worked out in many publications by the late German Indo-Europeanist Hans Krahe who coined the term ‘Old European’ for this particular stratum.][...such a pre-dialectal linguistic situation existed in central , northern and western Europe around 1500 BC…]
If so a river network of I-E names existed around 1500 BC, we can suppose they were named already a little sooner.
2- Surrounding the attested Italic and Celtic languages we found traces of other I-E ones with slight differences but also a lot of similitudes as Lusitanian, Ligurian (elite tongue?), Venetic, I-E Rhaetian (not the Etrusquic one), North-West Block I-E (some Belgaeic dialects?) ; all of them show a different set of differences and similitudes with Celtic and Italic dialects.

Moesan said...

3- What have we in West for culture evident change around 1500 BC and a bit before, if it isn’t northern BB culture ? CWC could have spoken an IE language if we consider the proposed ‘proto-Satem’ substratum in Saami dialects (curiously CWC could have send it through Denmark as well than through Finland, but the route through Finland seems a better bet).
4 - I hesitated long enough time, thinking first waves of BB could have been Basque speakers. The more eastern BB’s could have been indo-europeanized later at Bronze ? But what overwhelming colonisation have we in western Iberia at BA or at IA to explain Lusitanian supposed to be older than Celtic ? What links shows the Irish Gaelic phonetical evolution with Basque which could sustain the idea of a common Basquic substratum dating from BB’s time? Here I speak of language, not of DNA substratum. We see some Basquic and Gasconic influences on Castillan Romance; I don’t see convergence in it with evolution of Celtic into Irish Gaelic.
5- No certainty. But the hydronyms of Britain do support an I-E speaking presence before IA, and even before genuine plain MBA. What move of importance took place there outside BB’s?
Maybe the Y-R1b-DF27 question and Basque speaking people with ‘steppes’ DNA would imply an
I-E-ation of northeastern R1b-P312 by Y-R1a CWC people which didn’t concern southern R1b-P312 (so southern BB’s) ? But I don’t see anything in the linguistic field which seems confirming this and the explanation by a language shift of males Y-P312 incorporated in a Basque society and followed by a founding effect seems more realistic. This drift could have occurred among a pop with very small Y-R1b fringe with some punctual selective advantage among a too much endogam Basque group? Pure speculation of mine! So a lot of thing seems converging towards a vaste rather undifferentiated group of I-E dialects among all BB’s.

Moesan said...

Concerning the increase in EEF (and WHG) level begun already since MBA, it could be the result of pre-BB’s groups (megalithic Long Barrows people) reemerging or of progressive introgression of continental people based on trade and then exchanges. I have not the complete paper at hand and I cannot say anything about a possible unbalanced sexe mediation. The graphs I saw don’t show big differences between males and females.The level of increase is not too brutal and keeps on during EIA, and ideally could have been produced by a small move. All this if we believe the change is produced by “pure pre-BB” newcomers. But if we imagine another pop with only a little more EEF in it, the newcomers can have been very more numerous, what the study seems saying (50%?). All the way, the Kent individuals every time seems more EEF as a whole, and this may show us the entrance gate. For dates, a possible EEF impulse in BA, not too evident. The most evident is between 800 and 500 BC, and Scotland seems less or not concerned.
I don’t know what confidence to have in G25 but the new element seems showing some kind of proximity to regions of Northern and North-eastern France IA (= North-Gallia + Belgia) and a bit further to La Tène and Hallstatt of France, Germany and Czechia. Bavaria BB’s, Czech Tumuli people and Knovitz don’t seem too far either.
The today continental average pops of northern, northwestern, northeastern France, and Belgium seems to me close enough to Hallstatt and La Tène people, spite a light Germanic later input.

Concerning Picts, the toponymic evidences for a P-dialect there are not a dream (I posted here already) and the name of the Epidii tribe in Argyll, or the individual name Epogeni for a Caledonian are not in contradiction to that. The ‘maqqo’/’meqqo’ question is far from simple: it could be in Gaelic oghams spite the personal Pictish name. ATW, nothing excludes a first wave of primary poorly evolved common Celtic, or a pre-Celtic I-E BB dialect before the coming of defined P-Celtic bearers. The presence of Qw- or P- dialects in neighbouring places doesn’t say to us if either they are born there or are born elsewhere and come again in contact after first forking. I doubt Qw- Latin / Qw- Celtic bilinguism could have send P- Celtic. BTW Gaelic and Brittonic have a common trend to lenition and spiration in middle of words, but the results are far to be same in each group.
Spite certain shifts can occur independently I ‘m tempted to see the Qw- / P- shift as an Urnfields period event, linked to an elite or to long exchanges around Hungary and spread as well weswards than eastwards and southeastwards; personal thought outside the mainstream opinion. This “mutation” are not too precisely dated but they seem appearing around EIA and a bit later.

Steppe said...

Hello CopperAxe, it's Alain, I can unfortunately no longer get in at Anthrogenica although I have reset my password and have also written to the administrators, to this day no response even from two different email addresses, I don't understand why? I would be grateful if you could help me!

Ric Hern said...

I think the Celts were victims of Holding on Religiously to their Customs while their enemy modified their tactics. Basically playing Soccer while their enemies started playing Rugby. This was the downfall of many throughout history and will be the downfall of many in the future.

Wise dragon said...

According to Ancient Roman historians, some native Britons were closely related to Germanic people while others to Iberians.

“Their [the Britons’] hair is not only naturally blond, but they also use artificial means to increase this natural quality of color. For they continually wash their hair with limewash and draw it back from the forehead to the crown and to the nape of their neck, with the result that their appearance resembles that of Satyrs or of Pans, for their hair is so thickened by this treatment that it differs in no way from a horse’s mane. The men of Britain are taller than the Celti, and not so yellow-haired, although their bodies are of looser build.

The following is an indication of their size: I myself, in Rome, saw mere lads towering as much as half a foot above the tallest people in the city, although they were bandy-legged and presented no fair lines anywhere else in their figure.”    

– Strabo


“As is so often the case among barbarians, it is difficult to say whether the men who first dwelt in Britain are indigenous or whether they came thither. We may deduce arguments from the fact that their physical characteristics differ. For the reddish-yellow hair of the inhabitants of Caledonia [Northern Britain or Alba], as well as their sturdy limbs, point to a German origin; the swarthy complexion and curly hair of the Silures [from South West Britain], together with their position opposite Gaul, make us believe that the Iberians in ancient times crossed over and seized these territories. Those who are nearest to the Gauls resemble them, whether from the persistence of heredity, or whether, since the lands stretch out opposite each other, the climate has given the same character to the individuals.”

– Tacitus

“The Silures [from South West Britain] have swarthy features and are usually born with curly black hair, but the inhabitants of Caledonia [Northern Britain or Alba] have reddish hair and large but flaccid bodies. They are like the Gauls or the Spaniards, according as they are opposite either nation.”     

– Jordanes

Rob said...

@ David

''Here are some of those England EIA samples that look rather continental to me in terms of fine scale recent ancestry.''

yeah thatnks I noticed those. Looking at their context, there is nothing too unusual about them in terms of status ( I did not quite read up on all of them though), apart from their isotopic distinctiveness.
I was never convinced by Celtic from BB hypothesis, but we now have to ask - how did these apparently regular people effect langauge shift to Celtic across Britain & Ireland. Even if we see a few Gaulish warriors fleeing Caesar, how would they have imparted change in their new lands ?

Davidski said...

I'm not sure if social status of these continental-like people can tell us anything about language change, because they may have been commoners in an already Celtic speaking land.

I27379 is very interesting, because he's obviously a Celt and he closely resembles these continental-like EIA English samples, but he's from the LIA, so not directly relevant to the initial Celtic expansion into Britain.

Davidski said...

Well, I just found out that Pictish samples that didn't make it into the paper are identical to English/Scottish EBA samples.

Andrzejewski said...

@Davidski “ Well, I just found out that Pictish samples that didn't make it into the paper are identical to English/Scottish EBA samples.”

Nice! Nice to see that the scientific data confirms my theory regarding Pictish being a Beaker-related, pre-Celtic IE language.

Andrzejewski said...

@wise dragon So according to you, Strabo, Jordanes and Tacitus, EEF rich tribes were swarthy whereas WSH (Germanic) shifted ones were fairer and lighter pigmented?

Rob said...

@ Andrze

''So according to you, Strabo, Jordanes and Tacitus, EEF rich tribes were swarthy whereas WSH (Germanic) shifted ones were fairer and lighter pigmented?''

Yeah genius, Tacitus & Strabo said EEF & WSH were swarthy ewhilst the steppe-rich Andrzejewski's were beautiful blonds, like yourself

«Oldest ‹Older   1 – 200 of 434   Newer› Newest»